Robust optimization scheduling method for virtual power plant participating in electricity market

Abstract

The invention provides a robust optimization scheduling method for participation of a virtual power plant in an electric power market, and the method comprises the following steps: constructing a mathematical model of a power exchange station, an energy storage unit, a diesel generator, the behavior of participation in the electric power market, and the voltage of a power distribution network under a framework of the virtual power plant; constructing an uncertain set representing photovoltaic power generation and power conversion demands by using the predicted value and the fluctuation value under the confidence interval; establishing a two-stage robust optimization model by taking virtual power plant profit maximization as an objective function; and decomposing the two-stage robust optimization model into a main problem, a sub-problem outer layer and a sub-problem inner layer by adopting a nested column and constraint generation algorithm, adding the sub-problem inner layer into the constraint of the sub-problem outer layer by adopting a Kuhn-Tuck condition, and performing iterative solution through external and internal loops to obtain a scheduling scheme of the virtual power plant. According to the method, the income and the regulation and control capability of the virtual power plant in a complex power market environment are remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power market resource scheduling, and in particular relates to a robust optimization scheduling method for a virtual power plant to participate in a power market. Background Art

[0002] In recent years, with the rapid development of renewable energy and the increasing penetration of electric vehicles, the demand for flexibility in the power market has increased significantly. To this end, domestic and foreign scholars have extensively conducted research on the participation of electric vehicles in the power market, mainly focusing on the evaluation of adjustable capacity of electric vehicle aggregation, demand-side response, and vehicle-road-network interaction mechanisms. However, current research and practical applications still face many technical bottlenecks.

[0003] Among the charging modes of electric vehicles, plug-in mode is currently the most common method, but it has the following problems:

[0004] The plug-in mode usually uses DC or AC charging piles, which has a greater impact on the power grid, especially during peak load periods, which can easily aggravate the imbalance of the power grid;

[0005] In the vehicle-to-grid (V2G) mode, electric vehicles are highly random, and coupled with issues such as user privacy and communication delays, aggregate control is difficult.

[0006] In contrast, the battery swap mode, as a new charging method, completes energy interaction between the battery swap station and the power grid, which can not only meet the charging needs of users, but also has the following advantages:

[0007] The battery swap mode is more flexible and can complete the charging task in a shorter time;

[0008] The cost of battery replacement is low, and the batteries are centrally managed and maintained by battery replacement operators, which can provide timely feedback on battery status information, thereby reducing car owners' concerns about battery aging and safety.

[0009] Some studies have been conducted on the battery swapping model. For example:

[0010] A day-ahead optimal scheduling model for battery swap stations is proposed, focusing on the uncertainty of battery charging demand, time-of-use electricity prices, and constant current / constant voltage charging modes of batteries.

[0011] In the scenario where battery swap stations participate in the electricity market, some studies have proposed bidding strategies for the electric energy market and the frequency regulation ancillary service market;

[0012] For the heavy truck battery replacement scenario, the study proposed a daily scheduling and real-time correction strategy based on battery replacement demand, which can effectively achieve peak load shaving in the power grid.

[0013] However, the competitiveness of battery swap stations in the electricity spot market and ancillary service market is limited, mainly due to the following reasons:

[0014] The current scale and capacity of battery swap stations are not sufficient to meet the market's requirements for efficient operation alone;

[0015] Independently operated battery swap stations cannot fully utilize the synergistic effects of multiple energy resources.

[0016] To this end, incorporating battery swap stations into the unified management of virtual power plants and participating in the power market as part of power resources can significantly enhance their market competitiveness and flexibility. This approach can not only provide paid services to the power grid, but also achieve multi-energy coordinated scheduling under the framework of virtual power plants.

[0017] In terms of grid ancillary services, operating reserve is an important means to deal with the intermittency and volatility of renewable energy generation. Virtual power plants participate in grid operation through the power reserve market, and need to fully consider the uncertainty of internal resources (such as the uncertainty of renewable energy generation and the randomness of load). Existing research has proposed some solutions, such as:

[0018] The randomness of wind power and reserve calls is modeled using confidence intervals and confidence limits, and a multi-stage robust optimization model is solved using the Column and Constraint Generation (CCG) algorithm.

[0019] In a two-stage stochastic optimization framework, the uncertainty of electricity price and call is handled based on information gap decision theory.

[0020] Although the above research has made some progress, there are still the following shortcomings:

[0021] Current research mainly focuses on standby service constraints such as minimum standby amount or duration, but fails to adapt to more complex standby call patterns;

[0022] In more complex reserve services (such as the UK's STOR (Short Term Operating Reserve) service), in addition to the minimum aggregate capacity and duration, secondary response capabilities are also required, which places higher requirements on the optimal dispatch of virtual power plants;

[0023] Research on complex random call mechanisms is still limited, and most of them fail to fully consider the uncertainty coupling between renewable energy generation and reserve call. Summary of the invention

[0024] The purpose of the present invention is to address the deficiencies of the above-mentioned background technology, and to provide a robust optimization scheduling method for virtual power plants to participate in the electricity market, which incorporates battery swap stations, energy storage, diesel generators and photovoltaic power generation into the framework of the virtual power plant, and constructs an optimized scheduling scheme for participating in the electricity spot market and ancillary service market. This method fully considers the uncertainty of photovoltaic power generation and battery swapping demand through nested columns and constraint generation algorithms, and can achieve reliable scheduling of virtual power plants under complex random call mechanisms. Compared with the existing technology, the present invention significantly improves the revenue and regulation capabilities of virtual power plants in complex electricity market environments.

[0025] The technical solution adopted by the present invention is: a robust optimization scheduling method for a virtual power plant to participate in the power market, comprising the following steps:

[0026] In the framework of virtual power plant, mathematical models of battery swap stations, energy storage, diesel generators, electricity market participation and distribution network voltage are constructed, and the predicted values ​​and fluctuation values ​​under the confidence interval are used to construct the uncertainty set that characterizes the demand for photovoltaic power generation and battery swapping.

[0027] Among them, virtual power plants aggregate battery swap stations, energy storage, photovoltaics and diesel generators as power resources, and participate in the power spot and ancillary service markets as power trading entities;

[0028] Based on the above mathematical model, taking the profit maximization of virtual power plants as the objective function, a two-stage robust optimization model for virtual power plants to participate in the electricity market under multiple uncertainties is constructed; wherein the uncertainty set is used as a constraint in the optimization model to represent the uncertainty of photovoltaic power generation and battery replacement demand;

[0029] The two-stage robust optimization model is decomposed into the main problem, the outer layer of subproblems and the inner layer of subproblems by using the nested column and constraint generation algorithm. The inner layer of subproblems is added to the constraints of the outer layer of subproblems by using the Kuhn-Tucker condition, and the scheduling scheme of the virtual power plant is obtained by iteratively solving the problem through the outer and inner loops.

[0030] The scheduling plan includes: electricity sales allocation in different time periods; spare capacity allocation in different time periods; diesel generator start and stop status in different time periods; optimal charging and discharging strategies for energy storage and battery swap stations; and optimal values ​​for standby call, photovoltaic power generation and battery swap requirements.

[0031] In the above technical scheme, the mathematical model of the battery swap station is used to characterize: the limitations on the charging power and discharging power of the battery swap station; the battery swap station can only be in a charging, discharging or static state within any time step; the total energy stored in the battery swap station; the energy change of the battery swap station in each time step according to the charging and discharging operations, which complies with the charging and discharging efficiency; the relationship between the increase or decrease of energy of the battery swap station during the charging and discharging process and the number of batteries to be swapped in the battery swap station; the constraint relationship between the minimum number of batteries retained in the battery swap station and the energy of a single fully charged battery and the total energy of the battery swap station and the maximum rated energy of the generalized energy storage of the battery swap station; the initial energy of the battery swap station within the optimization period is consistent with the maximum rated energy of the generalized energy storage of the battery swap station.

[0032] In the above technical solution, the mathematical model of energy storage is used to characterize: the charging and discharging power limit of the energy storage system; the energy storage system can only be in a charging, discharging or static state in any time step; the energy of the energy storage system changes according to the charging and discharging operations in each time step, in accordance with the charging and discharging efficiency; the energy range constraint of the energy storage system; the energy at the end of the energy storage system is consistent with the initial state, which is half of the maximum rated energy of the energy storage.

[0033] In the above technical solution, the mathematical model of the diesel engine is used to characterize: the power generation constraint of the diesel generator; and the start-stop state constraint of the diesel generator.

[0034] In the above technical solution, the mathematical model of the behavior of participating in the electricity market is used to characterize that the sum of the power sales volume and the reserve call volume of the virtual power plant is equal to the net injected power of all equipment.

[0035] In the above technical solution, the mathematical model of the distribution network voltage is used to characterize: the power balance of each node; the flow of active power and reactive power in the lines between nodes; the voltage difference relationship between nodes; the relationship between the current flow of the line and the node voltage; and the range of constrained node voltage amplitudes.

[0036] In the above technical solution, the goal of the first stage of the optimization model is to maximize the revenue of the virtual power plant in the spot market and the reserve market minus the related costs;

[0037] The goal of the second phase is to minimize the system operating costs, including backup calls, diesel generator startup costs, energy storage and battery swap station operating costs, under uncertain conditions;

[0038] The decision variables in the first stage include the reserve capacity submitted to the reserve market, the amount of electricity sold by the virtual power plant to the spot market, and the start and stop status of the diesel generator;

[0039] The uncertainty decision variables in the second stage include: the grid's call for backup, the actual value of photovoltaic power generation, the actual value of the energy of the battery to be replaced, and the upward and downward adjustment coefficients of the average energy of photovoltaic power generation and the battery to be replaced;

[0040] In the second stage, the state decision variables of each device include: the power generation power of the diesel generator, the charging and discharging power of the energy storage, the charging and discharging state variables of the energy storage, the energy storage energy, the charging and discharging power of the battery swap station, the charging and discharging state variables of the battery swap station, the total energy of the battery swap station, and the node voltage.

[0041] In the above technical solution, the main problem is to maximize the total profit of the virtual power plant based on the decision variables in the first stage.

[0042] In the above technical solution, the outer layer of the sub-problem determines the worst case scenarios of photovoltaic, battery swap and backup calls based on the optimal value of the main problem and the charging and discharging binary variable scenario, and minimizes the total cost of the virtual power plant in the worst case.

[0043] In the above technical solution, the inner layer of the sub-problem determines the output of the battery swap station, energy storage, photovoltaic and diesel generators, and the charging and discharging binary variables of the battery swap station and energy storage system based on the worst case scenario obtained by the outer layer of the sub-problem, so as to maximize benefits or minimize costs.

[0044] The beneficial effects of the present invention are as follows: the present invention improves the flexibility of system scheduling by constructing a virtual power plant framework, aggregating multiple energy resources such as battery swap stations, energy storage, photovoltaic power generation and diesel generators. An uncertain set of photovoltaic power generation and battery swap requirements is constructed, and the optimization model is embedded in the form of constraints, which enhances the adaptability of the virtual power plant to fluctuations in new energy and changes in electricity demand. With the two-stage robust optimization model as the core, combined with the decomposition and solution of the main problem and sub-problems, the benefits and costs of the virtual power plant under uncertain conditions are optimized. The scheduling plan covers the distribution of electricity sales, spare capacity, start and stop status of diesel generators, and the charging and discharging strategies of energy storage and battery swap stations, providing executable optimization results for actual operation.

[0045] Furthermore, the mathematical model of the battery swap station of the present invention characterizes in detail the charging power limit, charging and discharging state and energy change of the battery swap station, ensuring that the operation of the battery swap station complies with physical constraints. By introducing the constraint relationship between the number of battery swap batteries and the energy of a single battery and the total energy, the charging and discharging strategy of the battery swap station can be dynamically adjusted to meet actual needs. The model takes into account the impact of charging and discharging efficiency, which helps to improve energy utilization efficiency. By setting the initial energy of the battery swap station to be consistent with the maximum rated energy, the continuity and stability of the battery swap station operation during the optimization period are guaranteed.

[0046] Furthermore, the mathematical model of energy storage of the present invention ensures that the energy storage operates within a safe range by limiting the charging and discharging power and state of the energy storage system. The model characterizes the energy changes of the energy storage system in each time step and its range constraints, supporting accurate dynamic energy management. Setting the energy at the end of the energy storage system to be consistent with the initial state provides long-term balance and controllability for energy storage. Considering the impact of charging and discharging efficiency on energy changes helps to improve the economy of the energy storage system.

[0047] Furthermore, the mathematical model of the diesel generator of the present invention ensures that the output power of the diesel generator meets physical and economic requirements through power generation constraints. The start-stop state constraints can dynamically adjust the operation mode of the diesel generator, reduce unnecessary startup and shutdown losses, and improve operation efficiency. As a backup power source, the diesel generator can provide stable power output in an emergency and improve the reliability of the system.

[0048] Furthermore, the mathematical model of the electricity market participation behavior of the present invention ensures the overall power balance of the virtual power plant by constraining the power sales and reserve call amount to be equal to the net injected power of all devices. The model can balance the participation strategies of the spot market and the reserve market and improve the benefits of the virtual power plant. The power output of all devices is coordinated within a framework to achieve the coordinated optimization of multiple energy sources.

[0049] Furthermore, the mathematical model of the distribution network voltage of the present invention ensures the voltage stability of the distribution network operation by constraining the voltage difference relationship between nodes and the node voltage amplitude range. It describes the relationship between the active and reactive power flow of the line, and provides an accurate power scheduling basis for the distribution network operation. The relationship constraint between the current flow and the node voltage can prevent line overload and improve the safety and reliability of the distribution network operation.

[0050] Furthermore, the two-stage robust optimization model of the present invention aims at maximizing revenue in the first stage and minimizing cost in the second stage, which clarifies the optimization direction and improves the overall benefits of the virtual power plant; covers multiple variables such as backup capacity, electricity sales, grid calls, equipment status, etc., and provides a comprehensive scheduling optimization solution; the optimization of uncertain variables (such as photovoltaic power generation and battery replacement demand) improves the ability of virtual power plants to cope with new energy fluctuations and load uncertainties.

[0051] Furthermore, the main problem of the present invention directly improves the economic benefits of the virtual power plant by optimizing the first-stage decision variables. The complex multi-stage problem is decomposed into the main problem and sub-problems for solution, which reduces the complexity of the solution. The spot market income and the reserve market income are considered at the same time, and the economic efficiency of the two is comprehensively balanced.

[0052] Furthermore, the outer layer of the sub-problem of the present invention aims to minimize the cost in the worst case, which improves the robustness of the virtual power plant scheduling scheme. By determining the worst case of photovoltaic, battery swap and standby call, the reliability and applicability of the scheduling strategy are optimized. Further optimization is performed based on the optimal value of the main problem, ensuring the coordination of problem decomposition and the effectiveness of the solution.

[0053] Furthermore, the present invention optimizes the worst case obtained from the outer layer of the sub-problem, determines the specific output of the battery swap station, energy storage, photovoltaic and diesel generators, and improves the accuracy of scheduling. The charging and discharging behavior of the battery swap station and energy storage system is optimized, and the adaptability of the virtual power plant to the fluctuation of new energy is enhanced. By dynamically optimizing the operating status of the equipment, the goal of maximizing benefits or minimizing costs is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 The overall framework for the virtual power plant of the present invention to participate in the electricity market;

[0055] Figure 2 A two-stage robust optimization scheduling framework for virtual power plants based on nested columns and constraint generation algorithms of the present invention;

[0056] Figure 3 This is a distribution diagram of a virtual power plant under a 6-node distribution network in the embodiment;

[0057] Figure 4 Generating an algorithm iterative convergence process for nested columns and constraints in an embodiment;

[0058] Figure 5 A diagram showing the reserve capacity, reserve call power and power sales of the virtual power plant in the embodiment;

[0059] Figure 6 The operating status of each asset of the virtual power plant under different uncertainty combinations in the embodiment;

[0060] Figure 7 The following is the daily voltage distribution diagram of the distribution network under different uncertainty combinations in the embodiment. DETAILED DESCRIPTION

[0061] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments to facilitate a clear understanding of the present invention, but they do not constitute a limitation on the present invention.

[0062] Example 1

[0063] like Figure 1 As shown, the present invention provides a robust optimization scheduling method for a virtual power plant to participate in an electricity market, comprising the following steps:

[0064] In the framework of virtual power plant, mathematical models of battery swap stations, energy storage, diesel generators, electricity market participation and distribution network voltage are constructed, and the predicted values ​​and fluctuation values ​​under the confidence interval are used to construct the uncertainty set that characterizes the demand for photovoltaic power generation and battery swapping.

[0065] Among them, virtual power plants aggregate battery swap stations, energy storage, photovoltaics and diesel generators as power resources, and participate in the power spot and ancillary service markets as power trading entities;

[0066] Based on the above mathematical model, taking the profit maximization of virtual power plants as the objective function, a two-stage robust optimization model for virtual power plants to participate in the electricity market under multiple uncertainties is constructed; wherein the uncertainty set is used as a constraint in the optimization model to represent the uncertainty of photovoltaic power generation and battery replacement demand;

[0067] The two-stage robust optimization model is decomposed into the main problem, the outer layer of subproblems and the inner layer of subproblems by using the nested column and constraint generation algorithm. The inner layer of subproblems is added to the constraints of the outer layer of subproblems by using the Kuhn-Tucker condition, and the scheduling scheme of the virtual power plant is obtained by iteratively solving the problem through the outer and inner loops.

[0068] The scheduling plan includes: electricity sales allocation in different time periods; spare capacity allocation in different time periods; diesel generator start and stop status in different time periods; optimal charging and discharging strategies for energy storage and battery swap stations; and optimal values ​​for standby call, photovoltaic power generation and battery swap requirements.

[0069] This embodiment specifically includes the following steps:

[0070] Step 1: Based on the short-term operating reserve requirements, construct a reserve call uncertainty set; use the predicted values ​​and fluctuation values ​​under the confidence interval to represent the uncertainty of battery replacement demand and photovoltaic power generation, and form a corresponding uncertainty set.

[0071] As shown in Table 1, short-term operating reserve requirements set minimum aggregate capacity, duration, and secondary response capability:

[0072] Table 1 UK STOR backup service technical indicators

[0073]

[0074] According to the specific technical indicators, the uncertainty of standby call is modeled, and the corresponding uncertainty set is obtained as follows:

[0075]

[0076] [d t ,d t+1 ,...,d t+n ]≥1-M d [1-(dt -d t-1 )] (1b)

[0077]

[0078] In the above formula, Λ d represents the uncertainty set of backup calls, n and m correspond to the duration of backup calls and the secondary response time, respectively, and d t is the grid call time. Formula (1b) constrains the duration of reserve call and uses the big M method to express the logical judgment relationship. When the reserve is not called at time t-1 and is called at time t, d t In the time range from t to t+n, it is 1. Formula (1c) constrains the secondary response time of the backup. When the call at time t-1 is 1 and the call ends at time t is 0, then d t At least one of the values ​​between time t and time t+m is 1, indicating that the virtual power plant has the ability to accept calls again during this period.

[0079] The uncertainty of battery replacement demand and photovoltaic power generation is expressed by the predicted value and fluctuation value under the confidence interval. r,s As shown below:

[0080]

[0081] In the above formula, Λ r,s represents the uncertain set of photovoltaic power generation and battery replacement demand, k t s+ and are the upward and downward adjustment coefficients of photovoltaic power generation and the average energy of the battery to be replaced, respectively, which are binary variables; P t r , P t r,～ and ΔP t r They are the actual value, predicted value and fluctuation value of photovoltaic power generation; and are the actual value, predicted value and fluctuation value of the battery energy; Γ r and Γ s are the uncertainties of photovoltaic power generation and battery replacement demand, respectively. Formulas (2b) and (2c) represent the actual values ​​of photovoltaic power generation and battery replacement demand in the worst case. Formula (2d) constrains that the upper and lower fluctuations cannot occur at the same time. Formula (2e) limits the size of uncertainty and reduces the conservatism of the optimization results. When Γ r =0 or Γ s =0, the actual value is equal to the predicted value. The larger the uncertainty coefficient, the greater the deviation of the actual value from the predicted value.

[0082] Step 2: Construct a framework for virtual power plants to participate in the electricity market, which incorporates battery swap stations, energy storage, diesel generators and photovoltaics as virtual power plant assets; mathematically model the battery swap stations, energy storage, diesel generators, electricity market participation behaviors and distribution network voltage models in turn.

[0083] The framework for virtual power plants to participate in the electricity market Figure 1 As shown in the figure, the virtual power plant aggregates resources such as battery swap stations, energy storage, photovoltaics and diesel generators, and participates in the electricity spot and ancillary service markets as an electricity trading entity. The virtual power plant incorporates battery swap stations into management, indirectly realizing the aggregation and regulation of electric vehicles; among them, diesel generators provide stable power output, and energy storage shifts the power generation of new energy in the time dimension. In addition to selling surplus electricity in the spot market, the virtual power plant provides backup services for the power grid through the ancillary service market. The backup capacity that can be provided must be determined in advance so that random calls within the day can be fully responded to.

[0084] The mathematical model of the battery swap station, energy storage, diesel generator, electricity market participation behavior and distribution network voltage model is:

[0085] (1) Mathematical model of electric vehicle battery swap station:

[0086]

[0087] In the formula, the subscript t represents the time, are the charging and discharging power of the battery swap station, is the charging and discharging state variable of the battery swap station, p s,max is the maximum power of the battery swap station, is the total energy of the battery swap station, The energy change caused by battery replacement, is the charging and discharging efficiency of the battery swap station, Δt is the charging and discharging time step, N min is the minimum number of batteries to be retained in the battery swap station, E max is the energy of a single fully charged battery, E s,max is the maximum rated energy of the generalized energy storage of the battery swap station, N t is the number of batteries that need to be replaced at present, is the average energy of these batteries, To optimize the initial energy of the battery swap station within the cycle. Formulas (3a)-(3d) are the charging and discharging power constraints of the battery swap station, and simultaneous charging and discharging are not allowed; Formula (3e) represents the generalized energy storage energy change of the battery swap station; Formula (3f) represents the minimum and maximum energy constraints of the battery swap station, requiring the battery swap station to retain a certain number of fully charged batteries; Formula (3g) represents the energy change caused by battery swapping, and Formula (3h) gives the initial energy of the battery swap station.

[0088] (2) Mathematical model of energy storage:

[0089]

[0090] Similar to the generalized energy storage of battery swap stations, are the charging and discharging power of energy storage, is the charge and discharge state variable of energy storage, p b,max To store energy and discharge the maximum power, For energy storage, is the charging and discharging efficiency of energy storage, E b,max is the maximum rated energy of energy storage, soc min 、soc max are the minimum and maximum states of charge of energy storage, respectively. To simulate the initial and final states of energy storage in the time domain, the energy at the final moment is required to be consistent with the initial state.

[0091] (3) Diesel generator mathematical model:

[0092]

[0093]

[0094] In the above formula, and They are the power generation capacity, minimum power generation capacity and maximum power generation capacity of the diesel generator respectively; Indicates the start and stop state of the diesel generator; the large M constant is denoted as MG. Formula (5a) indicates that the power output of the diesel generator is maintained within a given range after it is started; Formula (5b) indicates that after the diesel generator is started, it must remain in the start state for at least one more step, and Formula (5c) constrains the shutdown duration.

[0095] (4) Virtual power plants participate in power balance in the electricity market:

[0096]

[0097] Considering that the virtual power plant participates in both the electricity spot market and the reserve service market, formula (6) gives the power balance of the virtual power plant participating in the electricity market, that is, the sum of the power sales and reserve call of the virtual power plant is equal to the net injection power of all equipment. The amount of electricity sold by the virtual power plant to the electricity spot market; R t is the reserve capacity submitted to the reserve market a day ago, d t is the daily grid’s call for reserve, assuming d t ∈{0,1}, that is, once the grid reserves are called, they are fully used.

[0098] (5) Distribution network voltage model:

[0099]

[0100] In the above formula, if there is no diesel engine at busbar i, then The same goes for the power of energy storage and battery swap stations; is the active and reactive power of the load at bus i; N is the bus set of the distribution network, N j is the set of buses connected to bus j; p ij,t ,q ij,t and I ij are the active and reactive power flows and the square of the current amplitude of line ij, respectively, i,t is the square of the node voltage amplitude, α ij is the convexity coefficient, V min and V max are the minimum and maximum voltage amplitudes allowed. When virtual power plants participate in the electricity spot and reserve markets, the voltage constraints of the distribution network need to be considered. Equation (7a) represents the injected power of the node, (7b)-(7f) are the power flow models of the radial distribution network, and Equation (7e) gives the relationship between the convex voltage and current; Equation (7f) ensures that the scheduling of the virtual power plant will not cause the voltage of its supporting grid to exceed the limit.

[0101] Step 3: Construct a two-stage robust optimization model for virtual power plants to participate in the power market under multiple uncertainties. The optimization model takes virtual power plant profit maximization as the objective function, and the constraints include the operation constraints of various power resources, power balance constraints, and distribution network voltage constraints. Under multiple uncertainties, the first stage of the two-stage robust optimization model determines the day-ahead power sales, reserve capacity, and diesel generator start-stop status, and the second stage determines the operating output of each device under the known worst case. The model is as follows:

[0102]

[0103] coefficient and They correspond to standby income, electricity sales income, diesel generator startup cost, standby call income, diesel generator operating cost, energy storage operating cost and battery swap station operating cost respectively; is the relative start / stop state of the diesel generator at time t. The objective function is expressed as the profit maximization of the virtual power plant participating in the electricity spot and reserve market under multiple uncertainties. The constraints are the operation constraints of each asset within the virtual power plant, the start / stop expression of the diesel generator, and the maximum reserve and electricity sales constraints that the virtual power plant can provide. X is the set of decision variables in the first stage, U is the set of uncertainty decision variables in the second stage, and Y is the set of state decision variables of each device in the second stage.

[0104] Step 4: Use the nested column and constraint generation algorithm to decompose the two-stage robust optimization model into the main problem, the outer subproblem layer and the inner subproblem layer. Use the Kuhn-Tucker condition (Karush-Kuhn-Tucker, KKT) to add the inner subproblem layer to the constraints of the outer subproblem layer, and iterate the solution through outer and inner loops.

[0105] In step 4, the CCG algorithm framework is used to decompose the two-stage robust model into a main problem and sub-problems and solve them iteratively. Considering that the inner layer of the sub-problem contains binary variables of energy storage and battery swap stations, it is impossible to use the Kuhn-Tucker condition (Karush-Kuhn-Tucker, KKT) to convert the double-layer sub-problem into a single-layer problem. Therefore, a nested column and constraint generation algorithm is used here, which gives the upper and lower bounds of the optimal value through the main problem and sub-problem respectively. Constraints and decision variables under specific scenarios are continuously added to the main problem. The main problem and sub-problems are iteratively optimized until the upper and lower bounds converge to approximate values. The nested column and constraint generation algorithm divides the sub-problem into an outer layer and an inner layer. The binary variables of the inner layer are represented as scenarios in the outer layer of the sub-problem. The KKT condition is then used to convert the inner layer of the sub-problem into constraints of the outer layer, and then the inner and outer layers are iteratively solved. The framework of the nested column and constraint generation algorithm can be found in Figure 2 .

[0106] (1) Main problem model:

[0107]

[0108]

[0109] (3b), (3c), (6c), (6d)(9b)

[0110]

[0111]

[0112] ,

[0113] It is the decision variable set of the main problem; the scale of the main problem increases with the number of iterations k, and the worst case of the sub-problem is approached through the auxiliary variable θ of the cutting plane. Among them, They are the optimal values ​​of call, photovoltaic power generation and battery replacement demand obtained by solving the subproblem k' times. Wherein, the superscript or subscript k' indicates the value represented by the corresponding character obtained by solving the subproblem k' times.

[0114] (2) Outer layer of sub-problems:

[0115]

[0116] KKT (Formula 11a-11f) (10c)

[0117]

[0118]

[0119]

[0120] , which is the set of decision variables in the outer layer of the subproblem. The size of the outer layer of the subproblem increases with the number of iterations kk, and the worst case of the inner layer of the subproblem is approached through continuous iteration through the auxiliary variable η of the cutting plane. Among them, They are the optimal values ​​of the binary variables of the charging and discharging states of the energy storage and the battery swap station obtained by iteratively solving the inner layer of the sub-problem kk'. Wherein, the superscript or subscript kk' indicates the value represented by the corresponding character obtained by iteratively solving the inner layer of the sub-problem kk'.

[0121] (3) Inner layer of sub-problems:

[0122]

[0123] In the above formula, and The power sales, reserve capacity and start / stop status of the diesel generator obtained in the kth iteration of the main problem are: and The optimal values ​​of standby call, photovoltaic power generation and battery swapping demand obtained in the kkth iteration of the outer layer of the subproblem are used as known quantities in the inner layer of the subproblem. Wherein, the superscript or subscript kk represents the value represented by the corresponding character obtained by solving the kkth iteration of the outer layer of the subproblem.

[0124] Based on the decomposed main problem, outer subproblem layer, and inner subproblem layer, the specific process of the nested column and constraint generation algorithm is as follows:

[0125] (1) Set the inner and outer robust bounds, convergence accuracy and iteration coefficients of the main problem and subproblems. Outer loop: Set the initial upper and lower bounds (UB, LB) UB = +∞, LB = -∞, Δ = 0.05, k = 1; Inner loop: inUB = +∞, inLB = -∞, inΔ = 0.05, kk = 1. Main problem scenario set Sub-questions

[0126] (2) Given the main problem of photovoltaic power generation Battery replacement demand and the initial worst-case scenario of the call Given the energy storage in the subproblem and battery swap stations Initial scenario for the binary variables of charge and discharge.

[0127] (3) Adding photovoltaic power generation to the main problem scenario set MPS Battery replacement demand and the initial worst-case scenario of the call

[0128] (4) Based on MPS, by optimizing the objective function of the main problem, the optimal first-stage decision variables such as power sales, backup capacity and diesel generator start and stop status are obtained. Specifically, the main problem MP(9a)-(9z) is solved, the objective function value is MP*, the outer loop upper limit UB=MP* is updated, and X is obtained in turn. *,k : Optimal electricity sales value Reserve and diesel generator start and stop values

[0129] (5) Adding energy storage to the sub-problem scenario set SPS and battery swap stations Scenario for binary variables.

[0130] (6) Results based on SPS and main problem X *,k , solve SMP(10a)-(10y), the objective function value is SMP*, update the inner loop lower limit inLB=SMP*, and get U *,kk : That is, generate the worst-case call backup in the outer objective function Photovoltaic power generation And the need for battery replacement worst-case scenario.

[0131] (7) The worst-case scenario result U based on the outer layer *,kk , through the KKT condition, the optimization problem in the inner layer of the sub-problem is simplified to solving the gradient condition and complementary relaxation condition of the Lagrangian function, optimizing the charging and discharging strategies of energy storage and battery swap stations in the inner layer of the sub-problem, and generating a binary variable scenario:

[0132] Solve SSP(11a)-(11f), the objective function value is SSP*, optimize the charging and discharging strategies of energy storage and battery swap stations in the inner layer of the sub-problem, and generate binary variable scenarios. Update the inner loop upper limit inUB=max(inUB, SSP*), and get the energy storage and battery swap stations Charge and discharge binary variables.

[0133] In the energy change formula of the energy storage system and the battery swap station, the KKT condition can help deal with the relationship between energy balance and charging and discharging power. The KKT condition ensures that the charging and discharging power of the battery swap station and energy storage will not exceed the maximum rated power; through the complementary relaxation condition, the KKT condition can ensure that the battery swap station and energy storage can only be in charging, discharging or static state in each time step.

[0134] By applying the KKT condition, the optimal solution of the inner layer of the subproblem (such as charging and discharging power, energy state, etc.) can generate a set of linearized constraints. These constraints are fed back to the outer layer of the subproblem (SMP) to adjust the worst-case scenario of the outer layer's uncertainty variables (such as photovoltaic power generation, battery replacement demand, etc.).

[0135] (8) If Then the subproblem iteration ends, θ * =(SMP*+SSP*) / 2, update the lower limit of the outer loop Otherwise, kk=kk+1, jump to step (5).

[0136] (9) If Then the algorithm ends; otherwise, jump to step (3), k=k+1.

[0137] Step 5: A simulation example is designed to verify the effectiveness of the robust control method of the virtual power plant in participating in the power reserve market and the spot market under multiple uncertainties.

[0138] In step 5, a simulation example of a 6-node distribution network is designed, taking into account the distribution network frame and voltage constraints, such as Figure 3 As shown, a diesel generator set is set at bus 1, a battery swap station is set at bus 2, and a photovoltaic and energy storage unit group is set at bus 4. Three uncertainty combinations are set in the simulation experiment, namely Tr, Te = {0,0}, {3,3}, {7,5}.

[0139] First, the convergence of the nested column and constraint generation algorithm is verified, such as Figure 4 As shown in the figure, both the inner and outer loops of the algorithm converge quickly within 2 iterations. Figure 5 The reserve capacity, power sales, and reserve call in the worst case that the virtual power plant can provide under different combinations of uncertainties are given. The reserve provided by the virtual power plant is greater than the lower limit of 0.5MW, meeting the random call requirements. Power sales and reserve calls are concentrated between 10:00 and 15:00. During this period, photovoltaic power generation is high and there is a large amount of surplus power in the virtual power plant. The reserve provided by the virtual power plant is both capped at 1.5MW from 0:00 to 15:00 and from 22:00 to 23:00. During the peak load period between 15:00 and 22:00, the reserve capacity varies with the uncertainty of photovoltaic power generation and power exchange. Figure 6 The output of each asset in the virtual power plant and the SoC changes of energy storage are given. The diesel generator provides stable power output throughout the day, the energy storage system mainly absorbs photovoltaic power generation, and the battery swap station participates in the virtual power plant's electricity sales and provides spare capacity while meeting the battery swap demand. Figure 7 The voltage of the distribution network is shown, and the voltage of each bus is controlled within a given range.

[0140] The final simulation results show that the proposed virtual power plant framework and the corresponding robust control method are real and effective. The simulation results obtained by this method meet the requirements of short-term operation reserve and can effectively promote the participation of virtual power plants in power market transactions.

[0141] Example 2

[0142] The present invention provides a robust optimization dispatching system for a virtual power plant to participate in an electricity market, comprising:

[0143] Mathematical model and uncertainty set construction module, under the framework of virtual power plant, constructs mathematical models of battery swap stations, energy storage, diesel generators, participation in electricity market behavior and distribution network voltage, and uses predicted values ​​and fluctuation values ​​under confidence intervals to construct uncertainty sets that characterize photovoltaic power generation and battery swap demand; among them, the virtual power plant aggregates battery swap stations, energy storage, photovoltaics and diesel generators as power resources, and participates in the power spot and ancillary service markets as a power trading entity;

[0144] The optimization model building module, based on the above mathematical model, takes the profit maximization of the virtual power plant as the objective function, and builds a two-stage robust optimization model for the virtual power plant to participate in the power market under multiple uncertainties; wherein the uncertainty set is used as a constraint in the optimization model to represent the uncertainty of photovoltaic power generation and battery replacement demand;

[0145] The optimization model solving module uses a nested column and constraint generation algorithm to decompose the two-stage robust optimization model into a main problem, an outer layer of subproblems, and an inner layer of subproblems. The inner layer of subproblems is added to the constraints of the outer layer of subproblems using the Kuhn-Tucker condition, and it is solved iteratively through outer and inner loops to obtain the scheduling plan of the virtual power plant.

[0146] The scheduling plan includes: electricity sales allocation in different time periods; spare capacity allocation in different time periods; diesel generator start and stop status in different time periods; optimal charging and discharging strategies for energy storage and battery swap stations; and optimal values ​​for standby call, photovoltaic power generation and battery swap requirements.

[0147] Example 3

[0148] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a robust optimization scheduling method for a virtual power plant to participate in an electricity market as described in the above technical solution.

[0149] Example 4

[0150] The present invention also provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to execute a robust optimization scheduling method for a virtual power plant to participate in the power market as described in the above technical solution.

[0151] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0152] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0153] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0154] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0155] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the present invention and the claims, all of which are within the protection of the present invention.

[0156] The contents not described in detail in this specification belong to the prior art known to professional and technical personnel in this field.

Claims

1. A robust optimization scheduling method for virtual power plants to participate in the power market, characterized by: The following steps are involved: In the framework of virtual power plant, mathematical models of battery swap stations, energy storage, diesel generators, electricity market participation and distribution network voltage are constructed, and the predicted values ​​and fluctuation values ​​under the confidence interval are used to construct the uncertainty set that characterizes the demand for photovoltaic power generation and battery swapping. Among them, virtual power plants aggregate battery swap stations, energy storage, photovoltaics and diesel generators as power resources, and participate in the power spot and ancillary service markets as power trading entities; Based on the above mathematical model, taking the profit maximization of virtual power plants as the objective function, a two-stage robust optimization model for virtual power plants to participate in the electricity market under multiple uncertainties is constructed; wherein the uncertainty set is used as a constraint in the optimization model to represent the uncertainty of photovoltaic power generation and battery replacement demand; The two-stage robust optimization model is decomposed into the main problem, the outer layer of subproblems and the inner layer of subproblems by using the nested column and constraint generation algorithm. The inner layer of subproblems is added to the constraints of the outer layer of subproblems by using the Kuhn-Tucker condition, and the scheduling scheme of the virtual power plant is obtained by iteratively solving the problem through the outer and inner loops. The scheduling plan includes: electricity sales allocation in different time periods; spare capacity allocation in different time periods; diesel generator start and stop status in different time periods; optimal charging and discharging strategies for energy storage and battery swap stations; and optimal values ​​for standby call, photovoltaic power generation and battery swap requirements.

2. A method according to claim 1, characterized in that: The mathematical model of the battery swap station is used to characterize: the limitations on the charging and discharging power of the battery swap station; the battery swap station can only be in a charging, discharging or static state within any time step; the total energy stored in the battery swap station; the energy change of the battery swap station in each time step according to the charging and discharging operations, which complies with the charging and discharging efficiency; the relationship between the increase or decrease in energy of the battery swap station during the charging and discharging process and the number of batteries to be swapped in the battery swap station; the constraint relationship between the minimum number of batteries retained in the battery swap station and the energy of a single fully charged battery and the total energy of the battery swap station and the maximum rated energy of the generalized energy storage of the battery swap station; the initial energy of the battery swap station within the optimization period is consistent with the maximum rated energy of the generalized energy storage of the battery swap station.

3. A method according to claim 1, characterized in that: The mathematical model of energy storage is used to characterize: the energy storage system charge and discharge power limit; the energy storage system can only be in the charging, discharging or static state in any time step; the energy of the energy storage system changes according to the charging and discharging operations in each time step, in accordance with the charging and discharging efficiency; the energy range constraint of the energy storage system; The energy at the end of the energy storage system is consistent with the initial state, which is half of the maximum rated energy of the energy storage.

4. A method according to claim 1, characterized in that: The mathematical model of the diesel engine is used to characterize: the power generation constraint of the diesel generator; and the start-stop state constraint of the diesel generator.

5. A method according to claim 1, characterized in that: The mathematical model of the electricity market participation behavior is used to characterize that the sum of the power sales and reserve call-up of the virtual power plant is equal to the net injected power of all equipment.

6. A method according to claim 1, characterized in that: The mathematical model of the distribution network voltage is used to characterize: the power balance of each node; the active power and reactive power flow of the lines between nodes; the voltage difference relationship between nodes; the relationship between the current flow of the line and the node voltage; and the range of constrained node voltage amplitudes.

7. A method according to claim 1, characterized in that: The objective of the first stage of the optimization model is to maximize the revenue of the virtual power plant in the spot market and the reserve market minus the associated costs; The goal of the second phase is to minimize the system operating costs, including backup calls, diesel generator startup costs, energy storage and battery swap station operating costs, under uncertain conditions; The decision variables in the first stage include the reserve capacity submitted to the reserve market, the amount of electricity sold by the virtual power plant to the spot market, and the start and stop status of the diesel generator; The uncertainty decision variables in the second stage include: the grid's call for backup, the actual value of photovoltaic power generation, the actual value of the energy of the battery to be replaced, and the upward and downward adjustment coefficients of the average energy of photovoltaic power generation and the battery to be replaced; In the second stage, the state decision variables of each device include: the power generation power of the diesel generator, the charging and discharging power of the energy storage, the charging and discharging state variables of the energy storage, the energy storage energy, the charging and discharging power of the battery swap station, the charging and discharging state variables of the battery swap station, the total energy of the battery swap station, and the node voltage.

8. A method according to claim 7, characterized in that: The main problem is to maximize the total profit of the virtual power plant based on the first-stage decision variables.

9. A method according to claim 8, characterized in that: The outer layer of the sub-problem determines the worst case scenarios of photovoltaic, battery swap and backup calls based on the optimal value of the main problem and the charging and discharging binary variable scenario, and minimizes the total cost of the virtual power plant in the worst case.

10. A method according to claim 9, characterized in that: The inner layer of the sub-problem determines the output of the battery swap station, energy storage, photovoltaic and diesel generators, and the charging and discharging binary variables of the battery swap station and energy storage system based on the worst case scenario obtained by the outer layer of the sub-problem to maximize benefits or minimize costs.

Images