WO2024067521A2 - 风储虚拟电厂在线调度方法和装置 - Google Patents

风储虚拟电厂在线调度方法和装置 Download PDF

Info

Publication number
WO2024067521A2
WO2024067521A2 PCT/CN2023/121233 CN2023121233W WO2024067521A2 WO 2024067521 A2 WO2024067521 A2 WO 2024067521A2 CN 2023121233 W CN2023121233 W CN 2023121233W WO 2024067521 A2 WO2024067521 A2 WO 2024067521A2
Authority
WO
WIPO (PCT)
Prior art keywords
wind
power plant
period
storage virtual
virtual power
Prior art date
Application number
PCT/CN2023/121233
Other languages
English (en)
French (fr)
Inventor
吴启仁
梅生伟
刘建平
沈子奇
徐飞
陈来军
齐腾云
李亚静
Original Assignee
中国长江三峡集团有限公司
中国三峡新能源(集团)股份有限公司
清华大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中国长江三峡集团有限公司, 中国三峡新能源(集团)股份有限公司, 清华大学 filed Critical 中国长江三峡集团有限公司
Publication of WO2024067521A2 publication Critical patent/WO2024067521A2/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
    • G06Q10/06312—Adjustment or analysis of established resource schedule, e.g. resource or task levelling, or dynamic rescheduling
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/067—Enterprise or organisation modelling
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Energy or water supply
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks
    • H02J3/28—Arrangements for balancing of the load in a network by storage of energy
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks
    • H02J3/28—Arrangements for balancing of the load in a network by storage of energy
    • H02J3/32—Arrangements for balancing of the load in a network by storage of energy using batteries with converting means
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks
    • H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
    • H02J3/381—Dispersed generators
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks
    • H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
    • H02J3/46—Controlling of the sharing of output between the generators, converters, or transformers
    • H02J3/466—Scheduling the operation of the generators, e.g. connecting or disconnecting generators to meet a given demand
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
    • H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J2300/00—Systems for supplying or distributing electric power characterised by decentralized, dispersed, or local generation
    • H02J2300/20—The dispersed energy generation being of renewable origin
    • H02J2300/28—The renewable source being wind energy

Definitions

  • the present invention relates to the field of renewable energy power generation technology, and in particular to an online scheduling method and device for a wind energy storage virtual power plant.
  • a virtual power plant is a power coordination and management system that aggregates and coordinates the optimization of renewable energy and flexible resources through advanced information and communication technologies and software systems. It participates in the power market and power grid operation as a special power plant.
  • a wind-storage virtual power plant is a typical virtual power plant. Its wind power output capacity and load demand are random. It is difficult to realize the dispatch of a wind-storage virtual power plant system based on real-time predicted wind power output capacity and load demand.
  • wind-storage virtual power plants mostly construct Markov decision process (MDP) models that describe system dynamics through state transition probabilities to analyze multi-stage decision-making problems in real-time scheduling.
  • MDP Markov decision process
  • multi-stage decisions often produce dimensionality disasters, and conventional methods are difficult to use to solve MDP models.
  • Wind-storage virtual power plants also develop scheduling strategies based on neural dynamic programming, reinforcement learning and other methods, but the computational overhead is still large.
  • the present invention proposes an online scheduling method and device for a wind-storage virtual power plant, which describes the joint scheduling problem of wind-storage virtual power plants by formulating a rolling cycle MDP model.
  • the pre-calculation algorithm is used to solve the rolling cycle MDP model to obtain the real-time scheduling strategy of the wind-storage virtual power plant, which avoids the difficulty of solving the scheduling strategy due to dimensionality problems and reduces the computational overhead.
  • the present invention provides an online scheduling method for a wind-storage virtual power plant, the method comprising:
  • the target period and the multiple periods thereafter constitute the target rolling cycle
  • the pre-calculation algorithm and the underlying optimization sub-model of the rolling cycle MDP model solving the wind storage virtual power plant dispatching strategy for the target period;
  • the wind storage virtual power plant state includes: the energy storage battery charge state of the wind storage virtual power plant, the power grid exchange power and wind power output power in the previous period;
  • the wind storage virtual power plant prediction information includes: a wind power output power prediction value and a load prediction value of the wind storage virtual power plant;
  • the wind-storage virtual power plant dispatching strategy includes: the output power of the cogeneration equipment of the wind-storage virtual power plant, the grid exchange power and the energy storage battery operating power.
  • the rolling optimization sub-model is solved to obtain a basic feasible strategy consisting of the wind-storage virtual power plant dispatching strategies in the non-target period in the target rolling cycle, including:
  • a greedy algorithm is used to solve the rolling optimization sub-model to obtain a basic feasible strategy consisting of a wind-storage virtual power plant dispatching strategy for non-target periods in the target rolling cycle.
  • the construction process of the rolling optimization sub-model includes:
  • the first objective function is constructed with the minimum operating cost of the wind-storage virtual power plant as the goal;
  • the rolling optimization sub-model is generated based on the first objective function and its corresponding constraints.
  • c t (X(t), A(t)) is the operating cost of the wind-storage virtual power plant in period t
  • X(t) is the state of the wind-storage virtual power plant in period t
  • A(t) is the dispatching strategy of the wind-storage virtual power plant in period t
  • E(t) is the state of charge of the energy storage battery in the wind-storage virtual power plant in period t
  • p w (t) is the wind power output of the wind-storage virtual power plant in period t
  • p G (t-1) is the grid exchange power of the wind-storage virtual power plant in period (t-1)
  • p G (t) is the grid exchange power of the wind-storage virtual power plant in period t.
  • Line power is the natural gas consumption of the i-th cogeneration device in the wind-storage virtual power plant in period t
  • H i (t) is the heating load of the i-th cogeneration device in the wind-storage virtual power plant in period t
  • p D (t) is the load of the wind-storage virtual power plant in period t
  • ⁇ T is the duration of the period
  • ⁇ (t) is the power exchange price of the power grid
  • c is the natural gas usage price
  • a i is the slope of the linear function between the output power of the i-th cogeneration device in the wind-storage virtual power plant in period t and the natural gas consumption
  • b i is the intercept of the linear function between the output power of the i-th cogeneration device in the wind-storage virtual power plant in period t
  • the dispatching strategy of the wind-storage virtual power plant in the target period is solved based on the basic feasible strategy, the pre-calculation algorithm and the underlying optimization sub-model of the rolling cycle MDP model, including:
  • the operating cost of the wind-storage virtual power plant is input into the underlying optimization sub-model, and the pre-calculation algorithm is used to solve the underlying optimization sub-model to obtain the wind-storage virtual power plant scheduling strategy for the target period.
  • the construction process of the underlying optimization sub-model includes:
  • the wind storage virtual power plant operating cost in the target period is compared with the non-target
  • the second objective function is constructed by taking the minimum sum of the expected operating costs of the wind-storage virtual power plant in each period as the goal;
  • the underlying optimization sub-model is generated based on the second objective function and its corresponding constraints.
  • c t (X(t), A(t)) is the operation cost of the wind-storage virtual power plant in period t, is the operating cost of the wind storage virtual power plant corresponding to the wind storage virtual power plant dispatching strategy in period t 1
  • X(t) is the state of the wind storage virtual power plant in period t
  • A(t) is the dispatching strategy of the wind storage virtual power plant in period t
  • E(t) is the state of charge of the energy storage battery in the wind storage virtual power plant in period t
  • p w (t) is the wind power output of the wind storage virtual power plant in period t
  • p G (t-1) is the grid exchange power of the wind storage virtual power plant in period (t-1)
  • p G (t) is the grid exchange power of the wind storage virtual power plant in period t.
  • ⁇ T is the duration of the period
  • ⁇ (t) is the power exchange price of the power grid
  • c is the natural gas usage price
  • a i is the slope of the linear function between the output power of the i-th cogeneration device in the wind-storage virtual power plant and the natural gas consumption in period t
  • b i is the intercept of the linear function between the output power of the i-th cogeneration device in the wind-storage virtual power plant and the natural gas
  • the present invention provides an online dispatching device for a wind-storage virtual power plant, the device comprising:
  • a target rolling cycle determination module used to form a target rolling cycle with a target period and multiple periods thereafter;
  • the first optimization solution module is used to input the state of the wind-storage virtual power plant in the target period and the forecast information of the wind-storage virtual power plant in the non-target period in the target rolling cycle into the rolling optimization submodel of the pre-constructed rolling cycle MDP model, solve the rolling optimization submodel, and obtain a basic feasible strategy composed of the wind-storage virtual power plant dispatching strategy in the non-target period in the target rolling cycle;
  • a second optimization solution module is used to solve the wind storage virtual power plant dispatching strategy for the target period based on the basic feasible strategy, the pre-calculation algorithm and the underlying optimization sub-model of the rolling cycle MDP model;
  • the wind storage virtual power plant state includes: the energy storage battery charge state of the wind storage virtual power plant, the power grid exchange power and wind power output power in the previous period;
  • the wind storage virtual power plant prediction information includes: a wind power output power prediction value and a load prediction value of the wind storage virtual power plant;
  • the wind-storage virtual power plant dispatching strategy includes: the output power of the cogeneration equipment of the wind-storage virtual power plant, the grid exchange power and the energy storage battery operating power.
  • the present invention provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the online scheduling method for the wind-storage virtual power plant as described in the first aspect is implemented.
  • the present invention provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the online scheduling method for the wind-storage virtual power plant as described in the first aspect is implemented.
  • the present invention provides an online scheduling method and device for a wind-storage virtual power plant, which takes into account the randomness of wind power generation and the requirements of real-time scheduling, and formulates a rolling period MDP model including a rolling optimization submodel and an underlying optimization submodel to describe the joint scheduling problem of the wind-storage virtual power plant.
  • the rolling optimization submodel is solved to obtain the basic feasible strategy corresponding to the rolling period to which the target period belongs, and then the preview algorithm is used to solve the underlying optimization submodel on the basis of the basic feasible strategy to obtain the target period scheduling strategy.
  • the target period scheduling strategy obtained by the preview algorithm is inevitably better than the scheduling strategy inferred by the target period rolling optimization submodel, and the optimization performance is improved.
  • the preview algorithm can effectively solve the real-time scheduling problem of multiple energy sources and can avoid the dimensionality problem caused by excessive state and decision space, the difficulty of optimization solution and the computational overhead are significantly reduced.
  • FIG1 is a schematic diagram of the structure of a wind-storage virtual power plant provided by the present invention.
  • FIG2 is a schematic flow chart of an online dispatching method for a wind-storage virtual power plant provided by the present invention
  • FIG. 3 is a schematic structural diagram of an online dispatching device for a wind-storage virtual power plant provided by the present invention
  • FIG4 is a schematic diagram of the structure of an electronic device for implementing an online scheduling method for a wind-storage virtual power plant provided by the present invention.
  • FIG 1 The specific structural diagram of the wind storage virtual power plant studied in the present invention is shown in Figure 1, which includes wind turbines, energy storage batteries, combined heat and power (CHP) equipment, information control center and smart meters, wherein the wind storage virtual power plant is connected to the upstream power grid, and with the flexibility of energy storage control (energy storage batteries can be charged by wind turbines, and can also release electricity to the transmission grid through transmission lines), the dispatchability of the original wind power is improved, and the operation state of the wind storage virtual power plant is smoothed.
  • the output power of the combined heat and power (CHP) equipment, the grid exchange power (the exchange power between the wind storage virtual power plant and the upstream power grid) and the energy storage charging and discharging power in the wind storage virtual power plant are dispatched and controlled in real time.
  • the present invention provides an online scheduling method for a wind-storage virtual power plant, as shown in FIG2 , comprising:
  • the target time period is time period t
  • the target rolling cycle is time period t to time period t+T-1
  • T is the number of time periods included in the rolling cycle set in the present invention.
  • the wind-storage virtual power plant forecast information of the target rolling period is input into the rolling optimization submodel of the pre-built rolling cycle MDP model, and the rolling optimization submodel is solved to obtain a basic feasible strategy consisting of the wind-storage virtual power plant dispatch strategy for the non-target period in the target rolling period;
  • the present invention considers applying the preview algorithm to the multi-stage real-time scheduling of the wind-storage virtual power plant.
  • the core idea of the preview algorithm is that, for the target rolling cycle, first find a set of feasible solutions that meet the system safety constraints, that is, the feasible solutions from the target period (t period) to the T-1th period after the target period (t+T-1 period), and the feasible solutions from the next period of the target period (t+1 period) to the T-1th period after the target period (t+T-1 period) are called the basic feasible strategy; then, on the basis of the basic feasible strategy, estimate the impact of the decision of the target period on the optimization target of the entire rolling cycle, and backtrack and correct the feasible solution of the target period (t period) by optimizing the value of the target, so as to obtain the final implementation strategy of the target period (t period), which is called the preview strategy.
  • the optimization target of the rolling optimization sub-model is for a single period, and it realizes the rolling optimization solution of continuous period decisions.
  • the state X(t) is known by default, so the wind power forecast information and load forecast information of the target period (period t) are used, and the preview strategy of the target period (period t) can be obtained by using the preview strategy.
  • the system state X(t+1) of the time period after the target time period (t+1 time period) can be obtained, and then scroll to the next cycle (t+1 time period to t+T time period), and according to the feasible solution inference and preview from the time period after the target time period (t+1 time period) to the Tth time period after the target time period (t+T time period), the preview strategy of the time period after the target time period (t+1 time period) is obtained. And so on. Until the scheduling strategies for all time periods within the observation range are generated.
  • the wind storage virtual power plant state includes: the energy storage battery charge state of the wind storage virtual power plant, the power grid exchange power and wind power output power in the previous period;
  • the wind storage virtual power plant prediction information includes: a wind power output power prediction value and a load prediction value of the wind storage virtual power plant;
  • the wind-storage virtual power plant dispatching strategy includes: the output power of the cogeneration equipment of the wind-storage virtual power plant, the grid exchange power and the energy storage battery operating power.
  • the present invention provides an online scheduling method for a wind-storage virtual power plant, which takes into account the randomness of wind power generation and the requirements of real-time scheduling, and formulates a rolling period MDP model including a rolling optimization submodel and an underlying optimization submodel to describe the joint scheduling problem of the wind-storage virtual power plant.
  • the rolling optimization submodel is solved to obtain the basic feasible strategy corresponding to the rolling period to which the target period belongs, and then the preview algorithm is used to solve the underlying optimization submodel on the basis of the basic feasible strategy to obtain the target period scheduling strategy.
  • the target period scheduling strategy obtained by the preview algorithm is inevitably better than the scheduling strategy inferred by the target period rolling optimization submodel, and the optimization performance is improved.
  • the preview algorithm can effectively solve the real-time scheduling problem of multiple energy sources and can avoid the dimensionality problem caused by excessive state and decision space, the difficulty of optimization solution and the computational overhead are significantly reduced.
  • the rolling optimization sub-model is solved to obtain a basic feasible strategy consisting of a wind-storage virtual power plant dispatching strategy for a non-target period in the target rolling cycle, including:
  • a greedy algorithm is used to solve the rolling optimization sub-model to obtain a basic feasible strategy consisting of a wind-storage virtual power plant dispatching strategy for non-target periods in the target rolling cycle.
  • the present invention utilizes a greedy algorithm to perform local optimal selection in each time period, thereby converting the multi-stage real-time scheduling problem into a single-stage optimization problem, and sequentially obtains feasible solutions for the target time period (t time period) to the T-1th time period after the target time period (t+T-1 time period), and combines the feasible solutions from the next time period of the target time period (t+1 time period) to the T-1th time period after the target time period (t+T-1 time period) into a group of feasible strategies.
  • the construction process of the rolling optimization sub-model includes:
  • the first objective function is constructed with the minimum operating cost of the wind-storage virtual power plant as the goal;
  • the rolling optimization sub-model is generated based on the first objective function and its corresponding constraints.
  • MDP Markov decision process
  • the rolling optimization sub-model in the rolling cycle MDP model constructs the first objective function with the minimum operating cost of the wind storage virtual power plant as the goal;
  • the decision for period t includes three parts: CHP output power, grid exchange power, and energy storage battery operating power, which can be expressed as follows:
  • N is the total number of cogeneration equipment included in the wind-storage virtual power plant.
  • the cost function of period t is determined by the state and decision of the system, including the grid power exchange cost and the fuel cost of the CHP unit. Therefore, the cost function of the wind-storage virtual power plant is expressed as follows:
  • ⁇ (t) is the power exchange price of the power grid in period t
  • c is the natural gas usage price.
  • the safety constraints of the wind-storage virtual power plant system include:
  • E(t) represents the state of charge of the energy storage battery during period t
  • p B (t) represents the output power of the energy storage battery during period t
  • ⁇ T is the period length
  • ⁇ c and ⁇ d represent the efficiency of charging and discharging the energy storage battery, respectively.
  • p G (t) represents the exchange power between the virtual power plant and the grid
  • p w (t) represents the power generation power of the wind turbine
  • p D (t) represents the total power of the system load.
  • the upper and lower limits of the charging and discharging power of the energy storage battery are as follows:
  • E represents the lower limit of the state of charge of the energy storage battery, Indicates the upper limit of the energy storage battery's state of charge. Indicates the upper limit of battery charging and discharging power;
  • the exchange power between the wind power virtual power plant and the power grid has capacity limitations and ramp limits.
  • the corresponding constraints are as follows: is the upper limit of the power exchange capacity, and ⁇ is the ramp rate:
  • the constraints are a set of constraints limited to the t period, and the overall problem is a linear programming:
  • the first objective function constructs the power balance constraint:
  • the upper and lower limits of the charging and discharging power of the first energy storage battery are:
  • the upper and lower limits of the output power of the first cogeneration equipment are:
  • p B* (t) is the theoretical lower bound of the operating power of the energy storage battery in period t, is the theoretical upper limit of the operating power of the energy storage battery in period t
  • p G* (t) is the theoretical lower limit of the grid exchange power of the wind-storage virtual power plant in period t
  • the theoretical lower/upper bounds of each decision variable in period t are calculated based on the system state E(t) and pG (t-1) by integrating the original constraints.
  • the rolling optimization model constructed by the present invention takes the lowest operating cost as the goal and realizes the real-time scheduling of a single period to ensure the optimal economy.
  • the single-stage objective function actually adopted by the rolling optimization sub-model of the present invention can also take the minimum weighted value of the wind storage virtual power plant operation cost and the energy storage battery life loss as the goal. On the basis of the wind storage virtual power plant structure model, it adds the consideration of the energy storage operation life.
  • the first goal The function is as follows:
  • ⁇ (t)p G (t) ⁇ T corresponds to the power generation cost
  • takes into account the indirect impact of energy storage power on the service life of energy storage
  • ⁇ B is the influencing factor, which is determined by the construction cost and specific parameters of the energy storage system.
  • the underlying optimization sub-model based on the basic feasible strategy, the pre-calculation algorithm and the rolling cycle MDP model solves the wind storage virtual power plant dispatching strategy for the target period, including:
  • the operating cost of the wind-storage virtual power plant is input into the underlying optimization sub-model, and the pre-calculation algorithm is used to solve the underlying optimization sub-model to obtain the wind-storage virtual power plant scheduling strategy for the target period.
  • the working mode of the pre-algorithm is different from the dynamic programming used in solving the MDP.
  • Dynamic programming requires forward recursion or backward recursion, followed by backtracking to obtain the final result. Since the basic feasible strategy provides an approximate value for cost calculation, the corresponding pre-algorithm only needs one recursive step to solve, avoiding the dimensionality problem of recursion. Therefore, the pre-algorithm can effectively solve the multi-energy scheduling problem in real time, and can avoid the dimensionality problem caused by excessive state and decision space. In comparison, the performance and calculation time of the algorithm proposed in the present invention are better than the existing methods.
  • the construction process of the bottom layer optimization sub-model includes:
  • the underlying optimization sub-model is generated based on the second objective function and its corresponding constraints.
  • the objective function of the single-stage optimization problem is the single-step cost function in the underlying optimization model.
  • the minimized scheduling strategy is the preview strategy for period t, that is, This strategy implies the reuse of basic feasible strategies.
  • constraints constructed adaptively for the second objective function are as follows:
  • the present invention considers the real-time scheduling problem of the wind storage virtual power plant, and the goal is to obtain a set of scheduling strategies to meet the power load and heat demand in real time at the lowest cost.
  • a rolling period MDP is formulated to describe the joint scheduling problem.
  • the pre-calculation algorithm is used to solve the rolling horizontal MDP problem. Since the pre-calculation algorithm can effectively solve the multi-energy scheduling problem in real time and can avoid the dimensionality problem caused by excessive state and decision space, the performance and calculation time of the algorithm proposed in the present invention are better than the existing methods.
  • the present invention can effectively handle the situation when the decision space is large, so that a scheduling strategy with better cost can be obtained on the basis of ensuring real-time performance.
  • FIG3 illustrates a schematic diagram of the structure of the online dispatching of a wind storage virtual power plant. As shown in FIG3, the device includes:
  • a target rolling cycle determination module 21 is used to form a target rolling cycle with a target period and multiple periods thereafter;
  • the first optimization solution module 22 is used to input the wind storage virtual power plant state in the target period and the wind storage virtual power plant forecast information in the non-target period in the target rolling cycle into the rolling optimization sub-model of the pre-constructed rolling cycle MDP model, solve the rolling optimization sub-model, and obtain a basic feasible strategy composed of the wind storage virtual power plant dispatch strategy in the non-target period in the target rolling cycle;
  • the second optimization solution module 23 is used to solve the wind storage virtual power plant dispatching strategy in the target period based on the basic feasible strategy, the pre-calculation algorithm and the underlying optimization sub-model of the rolling cycle MDP model;
  • the wind storage virtual power plant state includes: the energy storage battery charge state of the wind storage virtual power plant, the power grid exchange power and wind power output power in the previous period;
  • the wind storage virtual power plant prediction information includes: a wind power output power prediction value and a load prediction value of the wind storage virtual power plant;
  • the wind-storage virtual power plant dispatching strategy includes: the output power of the cogeneration equipment of the wind-storage virtual power plant, the grid exchange power and the energy storage battery operating power.
  • the present invention provides an online dispatching device for a wind-storage virtual power plant, which takes into account the randomness of wind power generation and the requirements of real-time dispatching, and formulates a rolling period MDP model including a rolling optimization submodel and an underlying optimization submodel to describe the joint dispatching problem of the wind-storage virtual power plant.
  • the rolling optimization submodel is solved to obtain the basic feasible strategy corresponding to the rolling period to which the target period belongs, and then the preview algorithm is used to solve the underlying optimization submodel on the basis of the basic feasible strategy to obtain the target period dispatching strategy.
  • the target period dispatching strategy obtained by the preview algorithm is inevitably better than the dispatching strategy inferred by the rolling optimization submodel of the target period, and the optimization performance is improved.
  • the preview algorithm can effectively solve the real-time dispatching problem of multiple energy sources and can avoid the dimensionality problem caused by the excessive state and decision space, the difficulty of optimization solution and the computational overhead are significantly reduced.
  • the first optimization solution module 22 is specifically used for:
  • a greedy algorithm is used to solve the rolling optimization sub-model to obtain a basic feasible strategy consisting of a wind-storage virtual power plant dispatching strategy for non-target periods in the target rolling cycle.
  • the device further includes a construction module, and the construction module specifically includes a first construction unit for constructing a rolling optimization sub-model, and the first construction unit includes:
  • a first objective function construction subunit is used to construct a first objective function with the minimum operating cost of the wind-storage virtual power plant as the goal;
  • a constraint construction subunit of the first objective function used to construct a power balance constraint, upper and lower limit constraints on the charging and discharging power of the first energy storage battery, upper and lower limit constraints on the first power grid exchange power, and upper and lower limit constraints on the output power of the first cogeneration device for the first objective function;
  • the first generating sub-unit is used to generate the rolling optimization sub-model based on the first objective function and its corresponding constraints.
  • c t (X(t), A(t)) is the operating cost of the wind-storage virtual power plant in period t
  • X(t) is the state of the wind-storage virtual power plant in period t
  • A(t) is the dispatching strategy of the wind-storage virtual power plant in period t
  • E(t) is the state of charge of the energy storage battery in the wind-storage virtual power plant in period t
  • p w (t) is the wind power output of the wind-storage virtual power plant in period t
  • p G (t-1) is the grid exchange power of the wind-storage virtual power plant in period t-1
  • p G (t) is the grid exchange power of the wind-storage virtual power plant in period t.
  • ⁇ i is the output power of the i-th cogeneration device in the wind-storage virtual power plant during period t
  • p w (t) is the operating power of the energy storage battery in the wind-storage virtual power plant during period t
  • H i (t) is the heating load of the i-th cogeneration device in the wind-storage virtual power plant in period t
  • p D (t) is the load of the wind-storage virtual power plant in period t
  • ⁇ T is the duration of the period
  • ⁇ (t) is the power exchange price of the power grid
  • c is the natural gas usage price
  • a i is the output power and natural gas of the t-th cogeneration device in the wind-storage virtual power plant in period t
  • b i is the intercept of the linear function between the output power of the ith cogeneration device in the wind-storage virtual power plant and the natural gas consumption in period t
  • ⁇ c is the
  • the second optimization solution module 22 specifically includes:
  • a determination unit used to determine the wind storage virtual power plant operation cost corresponding to the wind storage virtual power plant dispatching strategy during the non-target period of the target rolling cycle
  • the preview algorithm optimization solving unit is used to input the operating cost of the wind-storage virtual power plant into the underlying optimization sub-model, use the preview algorithm to solve the underlying optimization sub-model, and obtain the wind-storage virtual power plant scheduling strategy for the target period.
  • the construction module includes a second construction unit for constructing an underlying optimization sub-model, and the second construction unit includes:
  • a second objective function construction subunit is used to construct a second objective function with the goal of minimizing the sum of the expected operating costs of the wind-storage virtual power plant in the target period and the wind-storage virtual power plant in the non-target period in the target rolling cycle;
  • the constraint construction subunit of the second objective function is used to construct a power balance constraint, a state of charge constraint of the energy storage battery, an upper and lower limit constraint of the charge and discharge power of the second energy storage battery, and a Second, the upper and lower limits of the power exchanged by the power grid and the upper and lower limits of the output power of the second cogeneration equipment;
  • the second generating sub-unit is used to generate the underlying optimization sub-model based on the second objective function and its corresponding constraints.
  • FIG4 illustrates a schematic diagram of the physical structure of an electronic device.
  • the electronic device may include: a processor 410, a communication interface 420, a memory 430 and a communication bus 440, wherein the processor 410, the communication interface 420 and the memory 430 communicate with each other through the communication bus 440.
  • the processor 410 may call the logic instructions in the memory 430 to execute the online scheduling method of the wind storage virtual power plant, which method includes: forming a target rolling cycle with a target period and multiple periods thereafter; inputting the state of the wind storage virtual power plant in the target period and the prediction information of the wind storage virtual power plant in the non-target period in the target rolling cycle into the rolling optimization submodel of the pre-constructed rolling cycle MDP model, solving the rolling optimization submodel, and obtaining a basic feasible strategy composed of the wind storage virtual power plant scheduling strategy for the non-target period in the target rolling cycle; based on the basic feasible strategy
  • the wind storage virtual power plant scheduling strategy for the target period is solved by using the pre-running algorithm and the underlying optimization sub-model of the rolling cycle MDP model; wherein the state of the wind storage virtual power plant includes: the state of charge of the energy storage battery of the wind storage virtual power plant, the grid exchange power and the wind power output power in the previous period; the prediction information of the wind storage virtual power plant includes: the wind power output power forecast value and the load forecast
  • the logic instructions in the above-mentioned memory 430 can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product.
  • the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art.
  • the computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention.
  • the aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.
  • the present invention further provides a computer program product, the computer program product comprising a computer program, the computer program can be stored in a non-transitory computer-readable storage medium, and when the computer program is executed by a processor, the computer can execute the wind energy storage virtual power generation system provided by the above methods.
  • a plant online scheduling method comprising: forming a target rolling cycle with a target period and multiple periods thereafter; inputting the state of a wind-storage virtual power plant in the target period and the prediction information of the wind-storage virtual power plant in non-target periods in the target rolling cycle into a rolling optimization submodel of a pre-constructed rolling cycle MDP model, solving the rolling optimization submodel, and obtaining a basic feasible strategy composed of the wind-storage virtual power plant scheduling strategy for non-target periods in the target rolling cycle; solving the wind-storage virtual power plant scheduling strategy for the target period based on the basic feasible strategy, the pre-running algorithm, and the underlying optimization submodel of the rolling cycle MDP model; wherein the state of the wind-storage virtual power plant comprises: the charge state of the energy storage battery of the wind-storage virtual power plant, the grid exchange power and the wind power output power in the previous period; the prediction information of the wind-storage virtual power plant comprises: the wind power output power prediction value and the load prediction value of the wind-storage virtual power plant; the
  • the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the wind-storage virtual power plant online scheduling method provided by the above-mentioned methods, the method comprising: forming a target rolling cycle with a target period and multiple periods thereafter; inputting the state of the wind-storage virtual power plant in the target period and the wind-storage virtual power plant forecast information in the non-target period in the target rolling cycle into a rolling optimization submodel of a pre-constructed rolling cycle MDP model, solving the rolling optimization submodel, and obtaining a wind-storage virtual power plant scheduling strategy group composed of non-target periods in the target rolling cycle.
  • a basic feasible strategy is formed; based on the basic feasible strategy, the preview algorithm and the underlying optimization sub-model of the rolling cycle MDP model, the wind storage virtual power plant dispatching strategy for the target time period is solved; wherein the state of the wind storage virtual power plant includes: the charge state of the energy storage battery of the wind storage virtual power plant, the grid exchange power and the wind power output power in the previous time period; the wind storage virtual power plant prediction information includes: the wind power output power prediction value and the load prediction value of the wind storage virtual power plant; the wind storage virtual power plant dispatching strategy includes: the output power of the cogeneration equipment of the wind storage virtual power plant, the grid exchange power and the energy storage battery operating power.
  • the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to implement this embodiment. The purpose of the scheme. Ordinary technicians in this field can understand and implement it without paying any creative work.
  • each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware.
  • the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM/RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Marketing (AREA)
  • Power Engineering (AREA)
  • General Physics & Mathematics (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Development Economics (AREA)
  • Quality & Reliability (AREA)
  • Game Theory and Decision Science (AREA)
  • Operations Research (AREA)
  • Educational Administration (AREA)
  • Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Water Supply & Treatment (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

本发明提供风储虚拟电厂在线调度方法和装置，将目标时段的风储虚拟电厂状态以及目标滚动周期中非目标时段的风储虚拟电厂风电出力功率预测信息和负荷预测信息输入预先构建的滚动周期MDP模型的滚动优化子模型中，求解得到由目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略；基于基本可行策略、预演算法以及滚动周期MDP模型的底层优化子模型，求解目标时段的风储虚拟电厂调度策略。本发明通过制定滚动周期MDP模型来描述风储虚拟电厂的联合调度问题，并采用预演算法作为滚动周期MDP模型的求解算法，使得求解风储虚拟电厂的实时调度策略时避免了由于维数问题导致的求解难度大的情况，还降低了计算开销。

Description

风储虚拟电厂在线调度方法和装置
本申请要求于2022年09月26日提交中国专利局、申请号为202211176980.1、申请名称为“风储虚拟电厂在线调度方法和装置”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本发明涉及新能源发电技术领域，尤其涉及风储虚拟电厂在线调度方法和装置。
背景技术
虚拟电厂是一种通过先进信息通信技术和软件系统，聚合并协调优化可再生能源和灵活资源的电源协调管理系统，它作为一个特殊电厂参与电力市场和电网运行。风储虚拟电厂是一种典型的虚拟电厂，其风电出力能力和负荷需求具有随机性，根据实时预测的风电出力能力和负荷需求实现风储虚拟电厂系统调度有一定困难。
目前，风储虚拟电厂多构建通过状态转移概率描述系统动态的马尔科夫决策过程(MDP)模型来分析实时调度过程中的多阶段决策问题。然而，由于转移概率难以估计、状态空间和决策空间维数过大等，多阶段决策常产生维数灾，常规方法难以用来求解MDP模型。风储虚拟电厂还基于神经动态规划、强化学习等方法展开调度策略，但计算开销仍较大。
因此，亟需提出一种高效的风储虚拟电厂调度策略。
发明内容
为了解决上述技术问题，本发明提出一种风储虚拟电厂在线调度方法和装置，通过制定滚动周期MDP模型来描述风储虚拟电厂的联合调度问题，使 用预演算法求解滚动周期MDP模型得到风储虚拟电厂的实时调度策略，避免了由于维数问题导致的调度策略求解难度大的问题，同时降低了计算开销。
第一方面，本发明提供一种风储虚拟电厂在线调度方法，所述方法包括：
以目标时段及其之后的多个时段构成目标滚动周期；
将目标时段的风储虚拟电厂状态以及所述目标滚动周期中非目标时段的风储虚拟电厂预测信息输入预先构建的滚动周期MDP模型的滚动优化子模型中，求解所述滚动优化子模型，得到由所述目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略；
基于所述基本可行策略、预演算法以及所述滚动周期MDP模型的底层优化子模型，求解所述目标时段的风储虚拟电厂调度策略；
其中，所述风储虚拟电厂状态，包括：所述风储虚拟电厂的储能电池荷电状态、上一个时段电网交换功率和风电出力功率；
所述风储虚拟电厂预测信息，包括：所述风储虚拟电厂的风电出力功率预测值和负荷预测值；
所述风储虚拟电厂调度策略，包括：所述风储虚拟电厂的热电联产设备输出功率、电网交换功率和储能电池运行功率。
根据本发明提供的风储虚拟电厂在线调度方法，所述求解所述滚动优化子模型，得到由所述目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略，包括：
采用贪心算法求解所述滚动优化子模型，得到由所述目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略。
根据本发明提供的风储虚拟电厂在线调度方法，所述滚动优化子模型的构建过程，包括：
以风储虚拟电厂最小运行成本为目标构建第一目标函数；
为所述第一目标函数构建功率平衡约束、第一储能电池充放电功率上下限约束、第一电网交换功率上下限约束和第一热电联产设备输出功率上下限约束；
基于所述第一目标函数及其对应的约束生成所述滚动优化子模型。
根据本发明提供的风储虚拟电厂在线调度方法，所述第一目标函数的表达式如下所示：

X(t)＝[E(t),pG(t-1),pw(t)]

所述第一目标函数构建功率平衡约束的表达式如下所示：
第一储能电池充放电功率上下限约束的表达式如下所示：


第一电网交换功率上下限约束的表达式如下所示：


第一热电联产设备输出功率上下限约束的表达式如下所示：


其中，ct(X(t),A(t))为t时段的风储虚拟电厂运行成本，X(t)为t时段的风储虚拟电厂状态，A(t)为t时段的风储虚拟电厂调度策略，E(t)为t时段下风储虚拟电厂中储能电池的荷电状态，pw(t)为t时段下所述风储虚拟电厂的风电出力功率，pG(t-1)为(t-1)时段下风储虚拟电厂的电网交换功率，pG(t)为t时段下风储虚拟电厂的电网交换功率，为t时段下风储虚拟电厂中第i个热电联产设备输出功率，pw(t)为t时段下风储虚拟电厂的储能电池运 行功率，为t时段下风储虚拟电厂中第i个热电联产设备的天然气用量，Hi(t)为t时段下风储虚拟电厂中第i个热电联产设备的供热负荷，pD(t)为t时段下风储虚拟电厂的负荷，ΔT为时段的时长，λ(t)为电网功率交换价格，c为天然气使用价格，ai为t时段下风储虚拟电厂中第i个热电联产设备输出功率与天然气用量之间的线性函数中的斜率，bi为t时段下风储虚拟电厂中第i个热电联产设备输出功率与天然气用量之间的线性函数中的截距，αc为储能电池充电效率，αd为储能电池放电效率，E为储能电池荷电状态的下限，为储能电池荷电状态的上限，为储能电池运行功率的上限，pB* (t)为t时段下储能电池运行功率的理论下界，为t时段下储能电池运行功率的理论上界，δ为爬坡率，pG 为功率交换容量下限，为功率交换容量上限，pG* (t)为t时段下风储虚拟电厂的电网交换功率的理论下界，为t时段下风储虚拟电厂的电网交换功率的理论上界，为t时段下风储虚拟电厂中第i个热电联产设备输出功率的理论下界，为t时段下风储虚拟电厂中第i个热电联产设备输出功率的理论上界，vi 为电能向热能转化比率的下限，电能向热能转化比率的上限，为风储虚拟电厂中第i个热电联产设备输出功率上限。
根据本发明提供的风储虚拟电厂在线调度方法，所述基于所述基本可行策略、预演算法以及所述滚动周期MDP模型的底层优化子模型，求解所述目标时段的风储虚拟电厂调度策略，包括：
确定所述目标滚动周期非目标时段的风储虚拟电厂调度策略对应的风储虚拟电厂运行成本；
将所述风储虚拟电厂运行成本输入底层优化子模型中，采用所述预演算法求解所述底层优化子模型，得到所述目标时段的风储虚拟电厂调度策略。
根据本发明提供的风储虚拟电厂在线调度方法，所述底层优化子模型的构建过程，包括：
以所述目标时段的风储虚拟电厂运行成本与所述目标滚动周期中非目标 时段的风储虚拟电厂运行成本的期望之间的加和最小为目标构建第二目标函数；
为所述第二目标函数构建功率平衡约束、储能电池荷电状态约束、第二储能电池充放电功率上下限约束、第二电网交换功率上下限约束和第二热电联产设备输出功率上下限约束；
基于所述第二目标函数及其对应的约束生成所述底层优化子模型。
根据本发明提供的风储虚拟电厂在线调度方法，所述第二目标函数的表达式如下所示：


X(t)＝[E(t),pG(t-1),pw(t)]

所述功率平衡约束的表达式如下所示：
所述储能电池荷电状态约束的表达式如下所示：
E(t+1)＝E(t)-pB(t)ΔTαc,pB(t)≤0
E(t+1)＝E(t)-pB(t)ΔT/αd,pB(t)>0
所述第二储能电池充放电功率上下限约束的表达式如下所示：
所述第二电网交换功率上下限约束的表达式如下所示：

|pG(t)-pG(t-1)|≤δ
所述第二热电联产设备输出功率上下限约束的表达式如下所示：

其中，ct(X(t),A(t))为t时段的风储虚拟电厂运行成本，为t1时段的风储虚拟电厂调度策略对应的风储虚拟电厂运行成本，X(t)为t时段的风储虚拟电厂状态，A(t)为t时段的风储虚拟电厂调度策略，E(t)为t时段下风储虚拟电厂中储能电池的荷电状态，pw(t)为t时段下所述风储虚拟电厂的风电出力功率，pG(t-1)为(t-1)时段下风储虚拟电厂的电网交换功率，pG(t)为t时段下风储虚拟电厂的电网交换功率，为t时段下风储虚拟电厂中第i个热电联产设备输出功率，pw(t)为t时段下风储虚拟电厂的储能电池运行功率，为t时段下风储虚拟电厂中第i个热电联产设备的天然气用量，Hi(t)为t时段下风储虚拟电厂中第i个热电联产设备的供热负荷，pD(t)为t时段下风储虚拟电厂的负荷，ΔT为时段的时长，λ(t)为电网功率交换价格，c为天然气使用价格，ai为t时段下风储虚拟电厂中第i个热电联产设备输出功率与天然气用量之间的线性函数中的斜率，bi为t时段下风储虚拟电厂中第i个热电联产设备输出功率与天然气用量之间的线性函数中的截距，αc为储能电池充电效率，αd为储能电池放电效率，E为储能电池荷电状态的下限，为储能电池荷电状态的上限，为储能电池运行功率的上限，δ为爬坡率，pG 为功率交换容量下限，为功率交换容量上限，vi 为电能向热能转化比率的下限，电能向热能转化比率的上限，为风储虚拟电厂中第i个热电联产设备输出功率下限。
第二方面，本发明提供一种风储虚拟电厂在线调度装置，所述装置包括：
目标滚动周期确定模块，用于以目标时段及其之后的多个时段构成目标滚动周期；
第一优化求解模块，用于将目标时段的风储虚拟电厂状态以及所述目标滚动周期中非目标时段的风储虚拟电厂预测信息输入预先构建的滚动周期MDP模型的滚动优化子模型中，求解所述滚动优化子模型，得到由所述目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略；
第二优化求解模块，用于基于所述基本可行策略、预演算法以及所述滚动周期MDP模型的底层优化子模型，求解所述目标时段的风储虚拟电厂调度策略；
其中，所述风储虚拟电厂状态，包括：所述风储虚拟电厂的储能电池荷电状态、上一个时段电网交换功率和风电出力功率；
所述风储虚拟电厂预测信息，包括：所述风储虚拟电厂的风电出力功率预测值和负荷预测值；
所述风储虚拟电厂调度策略，包括：所述风储虚拟电厂的热电联产设备输出功率、电网交换功率和储能电池运行功率。
第三方面，本发明提供一种电子设备，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，所述处理器执行所述程序时实现如第一方面所述风储虚拟电厂在线调度方法。
第四方面，本发明提供一种非暂态计算机可读存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时实现如第一方面所述风储虚拟电厂在线调度方法。
本发明提供的一种风储虚拟电厂在线调度方法和装置，考虑了风力发电的随机性以及实时调度的要求，制定包含滚动优化子模型和底层优化子模型的滚动周期MDP模型来描述风储虚拟电厂的联合调度问题，应用时求解滚动优化子模型得到目标时段所属滚动周期对应的基本可行策略，而后在基本可行策略的基础上采用预演算法求解底层优化子模型得到目标时段调度策略。据预演算法得到的目标时段调度策略必然优于目标时段滚动优化子模型推理的调度策略，优化性能得以提升。另外，由于预演算法能够有效地求解多能源实时调度问题，且能够避免状态和决策空间过大带来的维数问题，因此优化求解难度以及计算开销都得到了显着的降低。
附图说明
为了更清楚地说明本发明或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作一简单地介绍，显而易见地，下面描述 中的附图是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1是本发明提供的风储虚拟电厂的结构示意图；
图2是本发明提供的风储虚拟电厂在线调度方法的流程示意图；
图3是本发明提供的风储虚拟电厂在线调度装置的结构示意图；
图4是本发明提供的实现风储虚拟电厂在线调度方法的电子设备的结构示意图。
具体实施方式
为使本发明的目的、技术方案和优点更加清楚，下面将结合本发明中的附图，对本发明中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
本发明研究的风储虚拟电厂具体结构示意图如图1所示，包括风力发电机、储能电池、热电联产(CHP)设备、信息控制中心和智能电表，其中，风储虚拟电厂与上游电网连通，借助储能控制(储能电池可以通过风电机组充电，也可以通过输电线路将电能放至输电电网)的灵活性，提高原有风电的可调度性，平滑风储虚拟电厂运行状态。该风储虚拟电厂中热电联产(CHP)设备输出功率、电网交换功率(风储虚拟电厂与上游电网的交换功率)以及储能充放电功率以实时方式进行调度控制。
下面结合图1-图4描述本发明的风储虚拟电厂在线调度方法和装置。
第一方面，本发明提供一种风储虚拟电厂在线调度方法，如图2所示，包括：
S11、以目标时段及其之后的多个时段构成目标滚动周期；
假设目标时段为t时段，那么目标滚动周期为t时段至t+T-1时段，T为本发明设定的滚动周期包含的时段的个数。
S12、将目标时段的风储虚拟电厂状态以及所述目标滚动周期中非目标时 段的风储虚拟电厂预测信息输入预先构建的滚动周期MDP模型的滚动优化子模型中，求解所述滚动优化子模型，得到由所述目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略；
本发明考虑将预演算法利用到风储虚拟电厂多阶段实时调度中，预演算法的核心思想是，针对目标滚动周期，首先寻找一组满足系统安全约束的可行解，即目标时段(t时段)至目标时段往后数的第T-1个时段(t+T-1时段)的可行解，将目标时段的后一个时段(t+1时段)至目标时段往后数的第T-1个时段(t+T-1时段)的可行解称为基本可行策略；随后在基本可行策略的基础上，估计目标时段的决策对整个滚动周期优化目标的影响，通过优化目标的取值来回溯修正目标时段(t时段)的可行解，从而得到目标时段(t时段)的最终实施策略，这一策略称作预演策略。
因此，本发明先构建滚动优化子模型来求解预演算法所需的基本可行策略。具体来讲，当目标时段(t时段)的状态已知的情况下，利用滚动优化子模型可以求解得到目标时段(t时段)的决策，目标时段的后一个时段(t+1时段)至目标时段往后数的第T-1个时段(t+T-1时段)的可行解组成的基本可行策略πb＝(Ab,t+1,Ab,t+2,…,Ab,t+T-1)，在系统参数合理的情况下，换句话讲，滚动优化子模型的优化目标针对单时段，实现的是连续时段决策的滚动优化求解。
S13、基于所述基本可行策略、预演算法以及所述滚动周期MDP模型的底层优化子模型，求解所述目标时段的风储虚拟电厂调度策略；
使用预演算法求解目标时段(t时段)决策时，状态X(t)是默认已知的，故使用目标时段(t时段)的风力发电功率预测信息和负荷预测信息，利用预演策略可以获得目标时段(t时段)的预演策略该时段策略执行完毕后，当接收到目标时段的后一个时段(t+1时段)的风力发电功率预测信息和负荷预测信息，可以得到目标时段的后一个时段(t+1时段)的系统状态X(t+1)，然后滚动到下一周期(t+1时段至t+T时段)，根据对目标时段的后一个时段(t+1时段)到目标时段往后数的第T个时段(t+T时段)的可行解推断以及预演，获得目标时段的后一个时段(t+1时段)的预演策略以此类推， 直至观测范围内所有时段的调度策略生成为止。
其中，所述风储虚拟电厂状态，包括：所述风储虚拟电厂的储能电池荷电状态、上一个时段电网交换功率和风电出力功率；
所述风储虚拟电厂预测信息，包括：所述风储虚拟电厂的风电出力功率预测值和负荷预测值；
所述风储虚拟电厂调度策略，包括：所述风储虚拟电厂的热电联产设备输出功率、电网交换功率和储能电池运行功率。
本发明提供的一种风储虚拟电厂在线调度方法，考虑了风力发电的随机性以及实时调度的要求，制定包含滚动优化子模型和底层优化子模型的滚动周期MDP模型来描述风储虚拟电厂的联合调度问题，应用时求解滚动优化子模型得到目标时段所属滚动周期对应的基本可行策略，而后在基本可行策略的基础上采用预演算法求解底层优化子模型得到目标时段调度策略。据预演算法得到的目标时段调度策略必然优于目标时段滚动优化子模型推理的调度策略，优化性能得以提升。另外，由于预演算法能够有效地求解多能源实时调度问题，且能够避免状态和决策空间过大带来的维数问题，因此优化求解难度以及计算开销都得到了显着的降低。
在上述各实施例的基础上，作为一种可选的实施例，所述求解所述滚动优化子模型，得到由所述目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略，包括：
采用贪心算法求解所述滚动优化子模型，得到由所述目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略。
本发明利用贪心算法，在每个时段进行局部最优选择，进而将多阶段实时调度问题转化为单阶段优化问题，依次针对目标时段(t时段)至目标时段往后数的第T-1个时段(t+T-1时段)求得可行解，并将目标时段的后一个时段(t+1时段)至目标时段往后数的第T-1个时段(t+T-1时段)的可行解组成一组可行策略。
在上述各实施例的基础上，作为一种可选的实施例，所述滚动优化子模型的构建过程，包括：
以风储虚拟电厂最小运行成本为目标构建第一目标函数；
为所述第一目标函数构建功率平衡约束、第一储能电池充放电功率上下限约束、第一电网交换功率上下限约束和第一热电联产设备输出功率上下限约束；
基于所述第一目标函数及其对应的约束生成所述滚动优化子模型。
考虑到风力发电存在的不确定性和实时调度的要求，通常采用滚动周期多阶段决策的马尔科夫决策过程(Markov decision process，MDP)来描述风储虚拟电厂的调度问题。
其中，滚动周期MDP模型中的滚动优化子模型以风储虚拟电厂最小运行成本为目标构建第一目标函数；
风储虚拟电厂在t时段的状态，包含电池荷电状态、上一时段电网交换功率、风电出力功率三部分，由下式表示：
X(t)＝[E(t),pG(t-1),pw(t)]
t时段的决策，包含CHP输出功率、电网交换功率、储能电池运行功率三部分，由下式表示：
其中，N为风储虚拟电厂中包含的热电联产设备的总数。
t时段的成本函数由系统的状态和决策共同确定，包含电网功率交换成本和CHP单元的燃料成本。因此，风储虚拟电厂成本函数由下式表示：
其中，是第i个CHP天然气用量与输出功率之间的线性函数，ai和bi分别为线性函数斜率和截距，其具体数值取决于第i个CHP的实际参数，λ(t)为t时段的电网功率交换价格，c为天然气使用价格。
故而，所述第一目标函数的表达式如下所示：

X(t)＝[E(t),pG(t-1),pw(t)]

风储虚拟电厂系统安全约束包括：
储能电池动态约束：
E(t+1)＝E(t)-pB(t)ΔTαc,pB(t)≤0
E(t+1)＝E(t)-pB(t)ΔT/αd,pB(t)>0
其中，E(t)表示t时段储能电池的荷电状态，pB(t)表示t时段储能电池的输出功率，ΔT为时段长度，αc和αd分别表示储能电池充、放电的效率。
风电虚拟电厂的功率平衡约束：
其中，pG(t)表示虚拟电厂与电网的交换功率，表示第i个CHP单元的输出功率，pw(t)表示风电机组的发电功率，pD(t)表示系统负荷的总功率。
储能电池充放电功率的上、下限约束如下：
其中，E表示储能电池荷电状态的下限，表示储能电池荷电状态的上限，表示电池充放电功率的上限；
风电虚拟电厂与电网的交换功率存在容量限制和爬坡限制，对应约束如下，其中为功率交换容量的上限，δ为爬坡率：

|pG(t)-pG(t-1)|≤δ
CHP单元的运行约束为

其中，为CHP单元输出功率上限，vi 为电能向热能转化比率的下限，为电能向热能转化比率的上限，Hi(t)为CHP单元的供热负荷需求。
由于滚动优化子模型需要求解单阶段优化问题，约束条件则是限制在t时段内的约束集合，问题整体为线性规划：
因此，适应性的为第一目标函数构建如下约束：
所述第一目标函数构建功率平衡约束：
第一储能电池充放电功率上下限约束：


第一电网交换功率上下限约束：

pG* (t)＝max{pG(t-1)-δ，pG }
第一热电联产设备输出功率上下限约束：


pB* (t)为t时段下储能电池运行功率的理论下界，为t时段下储能电池运行功率的理论上界，pG* (t)为t时段下风储虚拟电厂的电网交换功率的理论下界，为t时段下风储虚拟电厂的电网交换功率的理论上界，为t时段下风储虚拟电厂中第i个热电联产设备输出功率的理论下界，为t时段下风储虚拟电厂中第i个热电联产设备输出功率的理论上界，t时段各个决策变量的理论下/上界是根据系统状态E(t)和pG(t-1)，整合原有约束条件计算而得。
本发明构建的滚动优化模型以最低运行成本为目标，实现对单时段的实时调度保证经济性的最优。
需要注意的是，本发明滚动优化子模型实际采用的单阶段目标函数还可以以风储虚拟电厂运行成本和储能电池寿命损耗的加权值最低为目标，其在风储虚拟电厂结构模型的基础上，增加了对储能运转寿命的考虑，第一目标 函数如下：
λ(t)pG(t)ΔT对应发电成本，对应天然气使用成本，πB|pB(t)|为考虑了储能功率对储能使用寿命带来的间接影响，πB为影响因子，由储能系统的建设成本和具体参数共同决定。
其约束条件与风储虚拟电厂最低运行成本为目标的第一目标函数一致。
在上述各实施例的基础上，作为一种可选的实施例，所述基于所述基本可行策略、预演算法以及所述滚动周期MDP模型的底层优化子模型，求解所述目标时段的风储虚拟电厂调度策略，包括：
确定所述目标滚动周期非目标时段的风储虚拟电厂调度策略对应的风储虚拟电厂运行成本；
将所述风储虚拟电厂运行成本输入底层优化子模型中，采用所述预演算法求解所述底层优化子模型，得到所述目标时段的风储虚拟电厂调度策略。
可以理解的是，开展预演算法时，要基于基本可行策略来估计从t+1时段到t+T-1时段的预计成本，再求解综合成本函数的最小化问题的结果，修正t时段的策略以作为预演策略。
本发明中，预演算法的工作方式与求解MDP时使用的动态规划不同。动态规划需要前向递归或后向递归，随后回溯以获得最终结果。预演算法由于基本可行策略提供了成本计算的近似值，相应的预演算法中只需要递推一步求解，避免了递归的维数问题。因此，预演算法能够有效地实时求解多能源调度问题，且能够避免状态和决策空间过大带来的维数问题，相比起来，本发明所提算法的性能和计算时间都优于已有的方法。
在上述各实施例的基础上，作为一种可选的实施例，所述底层优化子模型的构建过程，包括：
以所述目标时段的风储虚拟电厂运行成本与所述目标滚动周期中非目标时段的风储虚拟电厂运行成本的期望之间的加和最小为目标构建第二目标函数；
为所述第二目标函数构建功率平衡约束、储能电池荷电状态约束、第二 储能电池充放电功率上下限约束、第二电网交换功率上下限约束和第二热电联产设备输出功率上下限约束；
基于所述第二目标函数及其对应的约束生成所述底层优化子模型。
单阶段优化问题的目标函数是底层优化模型中的单步成本函数，第二目标函数(综合成本函数最小化)的表达式如下所示：


X(t)＝[E(t),pG(t-1),pw(t)]

使该成本最小化的调度策略即为t时段的预演策略，即 该策略隐含了对基本可行策略的复用。
适应性的为第二目标函数构建的约束如下：
所述功率平衡约束：
所述储能电池荷电状态约束的表达式如下所示：
E(t+1)＝E(t)-pB(t)ΔTαc,pB(t)≤0
E(t+1)＝E(t)-pB(t)ΔT/αd,pB(t)>0
所述第二储能电池充放电功率上下限约束的表达式如下所示：
所述第二电网交换功率上下限约束的表达式如下所示：

|pG(t)-pG(t-1)|≤δ
所述第二热电联产设备输出功率上下限约束的表达式如下所示：

其中，为t1时段的风储虚拟电厂调度策略对应的风储虚拟电厂运行成本。
本发明考虑风储虚拟电厂的实时调度问题，目标是获得一组调度策略，以最小的成本实时满足电力负荷和热量需求。考虑到该系统中风力发电不确定性的影响以及该联合调度问题的实时性要求，制定滚动周期MDP来描述该联合调度问题。然后采用预演算法求解滚动水平MDP问题。由于预演算法能够有效地实时求解多能源调度问题，且能够避免状态和决策空间过大带来的维数问题，故本发明所提算法的性能和计算时间都优于已有的方法。
即本发明能够有效处理决策空间较大时的情况，从而能够在保证实时性的基础上得到成本较优的调度策略。
下面对本发明提供的风储虚拟电厂在线调度装置进行描述，下文描述的风储虚拟电厂在线调度装置与上文描述的风储虚拟电厂在线调度方法可相互对应参照。图3示例了风储虚拟电厂在线调度的结构示意图，如图3所示，所述装置包括：
目标滚动周期确定模块21，用于以目标时段及其之后的多个时段构成目标滚动周期；
第一优化求解模块22，用于将目标时段的风储虚拟电厂状态以及所述目标滚动周期中非目标时段的风储虚拟电厂预测信息输入预先构建的滚动周期MDP模型的滚动优化子模型中，求解所述滚动优化子模型，得到由所述目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略；
第二优化求解模块23，用于基于所述基本可行策略、预演算法以及所述滚动周期MDP模型的底层优化子模型，求解所述目标时段的风储虚拟电厂调度策略；
其中，所述风储虚拟电厂状态，包括：所述风储虚拟电厂的储能电池荷电状态、上一个时段电网交换功率和风电出力功率；
所述风储虚拟电厂预测信息，包括：所述风储虚拟电厂的风电出力功率预测值和负荷预测值；
所述风储虚拟电厂调度策略，包括：所述风储虚拟电厂的热电联产设备输出功率、电网交换功率和储能电池运行功率。
本发明提供的一种风储虚拟电厂在线调度装置，考虑了风力发电的随机性以及实时调度的要求，制定包含滚动优化子模型和底层优化子模型的滚动周期MDP模型来描述风储虚拟电厂的联合调度问题，应用时求解滚动优化子模型得到目标时段所属滚动周期对应的基本可行策略，而后在基本可行策略的基础上采用预演算法求解底层优化子模型得到目标时段调度策略。据预演算法得到的目标时段调度策略必然优于目标时段滚动优化子模型推理的调度策略，优化性能得以提升。另外，由于预演算法能够有效地求解多能源实时调度问题，且能够避免状态和决策空间过大带来的维数问题，因此优化求解难度以及计算开销都得到了显着的降低。
在上述各实施例的基础上，作为一种可选的实施例，所述第一优化求解模块22具体用于：
采用贪心算法求解所述滚动优化子模型，得到由所述目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略。
在上述各实施例的基础上，作为一种可选的实施例，所述装置还包括构建模块，所述构建模块具体包括构建滚动优化子模型的第一构建单元，所述第一构建单元包括：
第一目标函数构建子单元，用于以风储虚拟电厂最小运行成本为目标构建第一目标函数；
第一目标函数的约束构建子单元，用于为所述第一目标函数构建功率平衡约束、第一储能电池充放电功率上下限约束、第一电网交换功率上下限约束和第一热电联产设备输出功率上下限约束；
第一生成子单元，用于基于所述第一目标函数及其对应的约束生成所述滚动优化子模型。
在上述各实施例的基础上，作为一种可选的实施例，所述第一目标函数的表达式如下所示：

X(t)＝[E(t),pG(t-1),pw(t)]

所述第一目标函数构建功率平衡约束的表达式如下所示：
第一储能电池充放电功率上下限约束的表达式如下所示：


第一电网交换功率上下限约束的表达式如下所示：

pG*(t)＝max{pG(t-1)-δ,pG }
第一热电联产设备输出功率上下限约束的表达式如下所示：


其中，ct(X(t),A(t))为t时段的风储虚拟电厂运行成本，X(t)为t时段的风储虚拟电厂状态，A(t)为t时段的风储虚拟电厂调度策略，E(t)为t时段下风储虚拟电厂中储能电池的荷电状态，pw(t)为t时段下所述风储虚拟电厂的风电出力功率，pG(t-1)为t-1时段下风储虚拟电厂的电网交换功率，pG(t)为t时段下风储虚拟电厂的电网交换功率，为t时段下风储虚拟电厂中第i个热电联产设备输出功率，pw(t)为t时段下风储虚拟电厂的储能电池运行功率，为t时段下风储虚拟电厂中第i个热电联产设备的天然气用量，Hi(t)为t时段下风储虚拟电厂中第i个热电联产设备的供热负荷，pD(t)为t时段下风储虚拟电厂的负荷，ΔT为时段的时长，λ(t)为电网功率交换价格，c为天然气使用价格，ai为t时段下风储虚拟电厂中第t个热电联产设备输出功率与天然气 用量之间的线性函数中的斜率，bi为t时段下风储虚拟电厂中第i个热电联产设备输出功率与天然气用量之间的线性函数中的截距，αc为储能电池充电效率，αd为储能电池放电效率，E为储能电池荷电状态的下限，为储能电池荷电状态的上限，为储能电池运行功率的上限，pB* (t)为t时段下储能电池运行功率的理论下界，为t时段下储能电池运行功率的理论上界，δ为爬坡率，pG 为功率交换容量下限，为功率交换容量上限，pG* (t)为t时段下风储虚拟电厂的电网交换功率的理论下界，为t时段下风储虚拟电厂的电网交换功率的理论上界，为t时段下风储虚拟电厂中第i个热电联产设备输出功率的理论下界，为t时段下风储虚拟电厂中第i个热电联产设备输出功率的理论上界，vi 为电能向热能转化比率的下限，电能向热能转化比率的上限，为风储虚拟电厂中第i个热电联产设备输出功率上限，N为风储虚拟电厂中包含的热电联产设备的总数。
在上述各实施例的基础上，作为一种可选的实施例，所述第二优化求解模块22，具体包括：
确定单元，用于确定所述目标滚动周期非目标时段的风储虚拟电厂调度策略对应的风储虚拟电厂运行成本；
预演算法优化求解单元，用于将所述风储虚拟电厂运行成本输入底层优化子模型中，采用所述预演算法求解所述底层优化子模型，得到所述目标时段的风储虚拟电厂调度策略。
在上述各实施例的基础上，作为一种可选的实施例，所述构建模块包括用于构建底层优化子模型的第二构建单元，所述第二构建单元包括：
第二目标函数构建子单元，用于以所述目标时段的风储虚拟电厂运行成本与所述目标滚动周期中非目标时段的风储虚拟电厂运行成本的期望之间的加和最小为目标构建第二目标函数；
第二目标函数的约束构建子单元，用于为所述第二目标函数构建功率平衡约束、储能电池荷电状态约束、第二储能电池充放电功率上下限约束、第 二电网交换功率上下限约束和第二热电联产设备输出功率上下限约束；
第二生成子单元，用于基于所述第二目标函数及其对应的约束生成所述底层优化子模型。
在上述各实施例的基础上，作为一种可选的实施例，所述第二目标函数的表达式如下所示：


X(t)＝[E(t),pG(t-1),pw(t)]

所述功率平衡约束的表达式如下所示：
所述储能电池荷电状态约束的表达式如下所示：
E(t+1)＝E(t)-pB(t)ΔTαc,pB(t)≤0
E(t+1)＝E(t)-pB(t)ΔT/αd,pB(t)>0
所述第二储能电池充放电功率上下限约束的表达式如下所示：
所述第二电网交换功率上下限约束的表达式如下所示：

|pG(t)-pG(t-1)|≤δ
所述第二热电联产设备输出功率上下限约束的表达式如下所示：

为t1时段的风储虚拟电厂调度策略对应的风储虚拟电厂运行成本。
第三方面，图4示例了一种电子设备的实体结构示意图，如图4所示，该电子设备可以包括：处理器(processor)410、通信接口(Communications Interface)420、存储器(memory)430和通信总线440，其中，处理器410，通信接口420，存储器430通过通信总线440完成相互间的通信。处理器410可以调用存储器430中的逻辑指令，以执行风储虚拟电厂在线调度方法，该方法包括：以目标时段及其之后的多个时段构成目标滚动周期；将目标时段的风储虚拟电厂状态以及所述目标滚动周期中非目标时段的风储虚拟电厂预测信息输入预先构建的滚动周期MDP模型的滚动优化子模型中，求解所述滚动优化子模型，得到由所述目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略；基于所述基本可行策略、预演算法以及所述滚动周期MDP模型的底层优化子模型，求解所述目标时段的风储虚拟电厂调度策略；其中，所述风储虚拟电厂状态，包括：所述风储虚拟电厂的储能电池荷电状态、上一个时段电网交换功率和风电出力功率；所述风储虚拟电厂预测信息，包括：所述风储虚拟电厂的风电出力功率预测值和负荷预测值；所述风储虚拟电厂调度策略，包括：所述风储虚拟电厂的热电联产设备输出功率、电网交换功率和储能电池运行功率。
此外，上述的存储器430中的逻辑指令可以通过软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本发明的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本发明各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。
第五方面，本发明还提供一种计算机程序产品，所述计算机程序产品包括计算机程序，计算机程序可存储在非暂态计算机可读存储介质上，所述计算机程序被处理器执行时，计算机能够执行上述各方法所提供的风储虚拟电 厂在线调度方法，该方法包括：以目标时段及其之后的多个时段构成目标滚动周期；将目标时段的风储虚拟电厂状态以及所述目标滚动周期中非目标时段的风储虚拟电厂预测信息输入预先构建的滚动周期MDP模型的滚动优化子模型中，求解所述滚动优化子模型，得到由所述目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略；基于所述基本可行策略、预演算法以及所述滚动周期MDP模型的底层优化子模型，求解所述目标时段的风储虚拟电厂调度策略；其中，所述风储虚拟电厂状态，包括：所述风储虚拟电厂的储能电池荷电状态、上一个时段电网交换功率和风电出力功率；所述风储虚拟电厂预测信息，包括：所述风储虚拟电厂的风电出力功率预测值和负荷预测值；所述风储虚拟电厂调度策略，包括：所述风储虚拟电厂的热电联产设备输出功率、电网交换功率和储能电池运行功率。
第六方面，本发明还提供一种非暂态计算机可读存储介质，其上存储有计算机程序，该计算机程序被处理器执行时实现以执行上述各方法提供的风储虚拟电厂在线调度方法，该方法包括：以目标时段及其之后的多个时段构成目标滚动周期；将目标时段的风储虚拟电厂状态以及所述目标滚动周期中非目标时段的风储虚拟电厂预测信息输入预先构建的滚动周期MDP模型的滚动优化子模型中，求解所述滚动优化子模型，得到由所述目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略；基于所述基本可行策略、预演算法以及所述滚动周期MDP模型的底层优化子模型，求解所述目标时段的风储虚拟电厂调度策略；其中，所述风储虚拟电厂状态，包括：所述风储虚拟电厂的储能电池荷电状态、上一个时段电网交换功率和风电出力功率；所述风储虚拟电厂预测信息，包括：所述风储虚拟电厂的风电出力功率预测值和负荷预测值；所述风储虚拟电厂调度策略，包括：所述风储虚拟电厂的热电联产设备输出功率、电网交换功率和储能电池运行功率。
以上所描述的装置实施例仅仅是示意性的，其中所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例 方案的目的。本领域普通技术人员在不付出创造性的劳动的情况下，即可以理解并实施。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到各实施方式可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件。基于这样的理解，上述技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品可以存储在计算机可读存储介质中，如ROM/RAM、磁碟、光盘等，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行各个实施例或者实施例的某些部分所述的方法。
最后应说明的是：以上实施例仅用以说明本发明的技术方案，而非对其限制；尽管参照前述实施例对本发明进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本发明各实施例技术方案的精神和范围。

Claims (10)

  1. 一种风储虚拟电厂在线调度方法，其特征在于，所述方法包括：
    以目标时段以及所述目标时段之后的多个时段构成目标滚动周期；
    将目标时段的风储虚拟电厂状态以及所述目标滚动周期中非目标时段的风储虚拟电厂预测信息输入预先构建的滚动周期MDP模型的滚动优化子模型中，求解所述滚动优化子模型，得到由所述目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略；
    基于所述基本可行策略、预演算法以及所述滚动周期MDP模型的底层优化子模型，求解所述目标时段的风储虚拟电厂调度策略；
    其中，所述风储虚拟电厂状态，包括：所述风储虚拟电厂的储能电池荷电状态、上一个时段电网交换功率和风电出力功率；
    所述风储虚拟电厂预测信息，包括：所述风储虚拟电厂的风电出力功率预测值和负荷预测值；
    所述风储虚拟电厂调度策略，包括：所述风储虚拟电厂的热电联产设备输出功率、电网交换功率和储能电池运行功率。
  2. 根据权利要求1所述的风储虚拟电厂在线调度方法，其特征在于，所述求解所述滚动优化子模型，得到由所述目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略，包括：
    采用贪心算法求解所述滚动优化子模型，得到由所述目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略。
  3. 根据权利要求1或2任一项所述的风储虚拟电厂在线调度方法，其特征在于，所述滚动优化子模型的构建过程，包括：
    以风储虚拟电厂最小运行成本为目标构建第一目标函数；
    为所述第一目标函数构建功率平衡约束、第一储能电池充放电功率上下限约束、第一电网交换功率上下限约束和第一热电联产设备输出功率上下限约束；
    基于所述第一目标函数及其对应的约束生成所述滚动优化子模型。
  4. 根据权利要求3所述的风储虚拟电厂在线调度方法，其特征在于，所 述第一目标函数的表达式如下所示：

    X(t)＝[E(t),pG(t-1),pw(t)]

    所述第一目标函数构建功率平衡约束的表达式如下所示：
    第一储能电池充放电功率上下限约束的表达式如下所示：


    第一电网交换功率上下限约束的表达式如下所示：


    第一热电联产设备输出功率上下限约束的表达式如下所示：


    其中，ct(X(t),A(t))为t时段的风储虚拟电厂运行成本，X(t)为t时段的风储虚拟电厂状态，A(t)为t时段的风储虚拟电厂调度策略，E(t)为t时段下风储虚拟电厂中储能电池的荷电状态，pw(t)为t时段下所述风储虚拟电厂的风电出力功率，pG(t-1)为(t-1)时段下风储虚拟电厂的电网交换功率，pG(t)为t时段下风储虚拟电厂的电网交换功率，为t时段下风储虚拟电厂中第i个热电联产设备输出功率，pw(t)为t时段下风储虚拟电厂的储能电池运行功率，为t时段下风储虚拟电厂中第i个热电联产设备的天然气用量， Hi(t)为t时段下风储虚拟电厂中第i个热电联产设备的供热负荷，pD(t)为t时段下风储虚拟电厂的负荷，ΔT为时段的时长，λ(t)为电网功率交换价格，c为天然气使用价格，ai为t时段下风储虚拟电厂中第i个热电联产设备输出功率与天然气用量之间的线性函数中的斜率，bi为t时段下风储虚拟电厂中第i个热电联产设备输出功率与天然气用量之间的线性函数中的截距，αc为储能电池充电效率，αd为储能电池放电效率，E为储能电池荷电状态的下限，为储能电池荷电状态的上限，为储能电池运行功率的上限，pB* (t)为t时段下储能电池运行功率的理论下界，为t时段下储能电池运行功率的理论上界，δ为爬坡率，pG 为功率交换容量下限，为功率交换容量上限，pG* (t)为t时段下风储虚拟电厂的电网交换功率的理论下界，为t时段下风储虚拟电厂的电网交换功率的理论上界，为t时段下风储虚拟电厂中第i个热电联产设备输出功率的理论下界，为t时段下风储虚拟电厂中第i个热电联产设备输出功率的理论上界，vi 为电能向热能转化比率的下限，电能向热能转化比率的上限，为风储虚拟电厂中第i个热电联产设备输出功率上限，N为风储虚拟电厂中包含的热电联产设备的总数。
  5. 根据权利要求1～3任一项所述的风储虚拟电厂在线调度方法，其特征在于，所述基于所述基本可行策略、预演算法以及所述滚动周期MDP模型的底层优化子模型，求解所述目标时段的风储虚拟电厂调度策略，包括：
    确定所述目标滚动周期非目标时段的风储虚拟电厂调度策略对应的风储虚拟电厂运行成本；
    将所述风储虚拟电厂运行成本输入底层优化子模型中，采用所述预演算法求解所述底层优化子模型，得到所述目标时段的风储虚拟电厂调度策略。
  6. 根据权利要求5所述的风储虚拟电厂在线调度方法，其特征在于，所述底层优化子模型的构建过程，包括：
    以所述目标时段的风储虚拟电厂运行成本与所述目标滚动周期中非目标 时段的风储虚拟电厂运行成本的期望之间的加和最小为目标构建第二目标函数；
    为所述第二目标函数构建功率平衡约束、储能电池荷电状态约束、第二储能电池充放电功率上下限约束、第二电网交换功率上下限约束和第二热电联产设备输出功率上下限约束；
    基于所述第二目标函数及其对应的约束生成所述底层优化子模型。
  7. 根据权利要求6所述的风储虚拟电厂在线调度方法，其特征在于，所述第二目标函数的表达式如下所示：


    X(t)＝[E(t),pG(t-1),pw(t)]

    所述功率平衡约束的表达式如下所示：
    所述储能电池荷电状态约束的表达式如下所示：
    E(t+1)＝E(t)-pB(t)ΔTαc,pB(t)≤0
    E(t+1)＝E(t)-pB(t)ΔT/αd,pB(t)>0
    所述第二储能电池充放电功率上下限约束的表达式如下所示：
    所述第二电网交换功率上下限约束的表达式如下所示：

    |pG(t)-pG(t-1)|≤δ
    所述第二热电联产设备输出功率上下限约束的表达式如下所示：

    其中，ct(X(t),A(t))为t时段的风储虚拟电厂运行成本，为t1时段的风储虚拟电厂调度策略对应的风储虚拟电厂运行成本，X(t)为t时段的风储虚拟电厂状态，A(t)为t时段的风储虚拟电厂调度策略，E(t)为t时段下风储虚拟电厂中储能电池的荷电状态，pw(t)为t时段下所述风储虚拟电厂的风电出力功率，pG(t-1)为(t-1)时段下风储虚拟电厂的电网交换功率，pG(t)为t时段下风储虚拟电厂的电网交换功率，为t时段下风储虚拟电厂中第i个热电联产设备输出功率，pw(t)为t时段下风储虚拟电厂的储能电池运行功率，为t时段下风储虚拟电厂中第i个热电联产设备的天然气用量，Hi(t)为t时段下风储虚拟电厂中第i个热电联产设备的供热负荷，pD(t)为t时段下风储虚拟电厂的负荷，ΔT为时段的时长，λ(t)为电网功率交换价格，c为天然气使用价格，ai为t时段下风储虚拟电厂中第i个热电联产设备输出功率与天然气用量之间的线性函数中的斜率，bi为t时段下风储虚拟电厂中第i个热电联产设备输出功率与天然气用量之间的线性函数中的截距，αc为储能电池充电效率，αd为储能电池放电效率，E为储能电池荷电状态的下限，为储能电池荷电状态的上限，为储能电池运行功率的上限，δ为爬坡率，pG 为功率交换容量下限，为功率交换容量上限，vi 为电能向热能转化比率的下限，电能向热能转化比率的上限，为风储虚拟电厂中第i个热电联产设备输出功率上限，N为风储虚拟电厂中包含的热电联产设备的总数。
  8. 一种风储虚拟电厂在线调度装置，其特征在于，所述装置包括：
    目标滚动周期确定模块，用于以目标时段以及所述目标时段之后的多个时段构成目标滚动周期；
    第一优化求解模块，用于将目标时段的风储虚拟电厂状态以及所述目标滚动周期中非目标时段的风储虚拟电厂预测信息输入预先构建的滚动周期MDP模型的滚动优化子模型中，求解所述滚动优化子模型，得到由所述目标滚动周期中非目标时段的风储虚拟电厂调度策略组成的基本可行策略；
    第二优化求解模块，用于基于所述基本可行策略、预演算法以及所述滚动周期MDP模型的底层优化子模型，求解所述目标时段的风储虚拟电厂调度策略；
    其中，所述风储虚拟电厂状态，包括：所述风储虚拟电厂的储能电池荷电状态、上一个时段电网交换功率和风电出力功率；
    所述风储虚拟电厂预测信息，包括：所述风储虚拟电厂的风电出力功率预测值和负荷预测值；
    所述风储虚拟电厂调度策略，包括：所述风储虚拟电厂的热电联产设备输出功率、电网交换功率和储能电池运行功率。
  9. 一种电子设备，包括存储器、处理器及存储在所述存储器上并可在所述处理器上运行的计算机程序，其特征在于，所述处理器执行所述程序时实现如权利要求1至7任一项所述风储虚拟电厂在线调度方法。
  10. 一种非暂态计算机可读存储介质，其上存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现如权利要求1至7任一项所述风储虚拟电厂在线调度方法。
PCT/CN2023/121233 2022-09-26 2023-09-25 风储虚拟电厂在线调度方法和装置 WO2024067521A2 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211176980.1 2022-09-26
CN202211176980.1A CN115423207A (zh) 2022-09-26 2022-09-26 风储虚拟电厂在线调度方法和装置

Publications (1)

Publication Number Publication Date
WO2024067521A2 true WO2024067521A2 (zh) 2024-04-04

Family

ID=84206188

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/121233 WO2024067521A2 (zh) 2022-09-26 2023-09-25 风储虚拟电厂在线调度方法和装置

Country Status (2)

Country Link
CN (1) CN115423207A (zh)
WO (1) WO2024067521A2 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118094965A (zh) * 2024-04-26 2024-05-28 国网浙江省电力有限公司经济技术研究院 一种虚拟电厂频率响应聚合等值方法、装置及存储介质
CN118134228A (zh) * 2024-05-10 2024-06-04 浙江大学 含电动汽车负荷的工业园区综合能源系统在线调度方法

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115940166B (zh) * 2023-03-10 2023-05-26 中国铁塔股份有限公司 基站调度方法、装置、电子设备及可读存储介质
CN116961188B (zh) * 2023-07-26 2024-05-10 中国铁塔股份有限公司 基站电源的调度方法、装置、电子设备和可读存储介质
CN117541030B (zh) * 2024-01-09 2024-04-26 中建科工集团有限公司 虚拟电厂优化运行方法、装置、设备及介质

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118094965A (zh) * 2024-04-26 2024-05-28 国网浙江省电力有限公司经济技术研究院 一种虚拟电厂频率响应聚合等值方法、装置及存储介质
CN118134228A (zh) * 2024-05-10 2024-06-04 浙江大学 含电动汽车负荷的工业园区综合能源系统在线调度方法

Also Published As

Publication number Publication date
CN115423207A (zh) 2022-12-02

Similar Documents

Publication Publication Date Title
WO2024067521A2 (zh) 风储虚拟电厂在线调度方法和装置
CN109636056A (zh) 一种基于多智能体技术的多能源微网去中心化优化调度方法
CN112800658A (zh) 一种考虑源储荷互动的主动配电网调度方法
CN113610311A (zh) 双层协同架构下考虑碳减排的综合能源服务商合作运行优化方法
CN112862253B (zh) 考虑风电不确定性的电-气-热储能协同规划方法及系统
JP2020524325A (ja) 環境コストおよびリアタイムの電気代の両方を考慮した変位可能な負荷モデルの構築方法
CN115954957A (zh) 一种复合发电系统的容量配置方法、装置和电子设备
CN116167483A (zh) 计及阶梯型需求响应的园区综合能源系统鲁棒调度方法
CN114881328A (zh) 计及气网混氢和低碳奖励的综合能源系统经济调度方法
CN112365034B (zh) 一种电热综合能源系统调度方法及系统
CN113128746A (zh) 一种基于多代理的综合能源系统多主体联合优化运行方法
CN111680816A (zh) 一种提供多重服务的储能系统运营方法及系统
CN115986833A (zh) 考虑两阶段需求响应的热电联供微网低碳经济调度方法
CN116826753A (zh) 基于粒子群算法的微电网双层协调优化调度方法及系统
CN115986743A (zh) 基于隐式决策法的电力跨区域配置优化方法、装置及设备
CN116777234A (zh) 一种两阶段鲁棒优化模型的构建、策略制定方法及系统
CN115600757A (zh) 海上风电共享储能参与现货市场交易协调优化方法及系统
CN115882460A (zh) 一种考虑需求侧管理的两阶段新能源微电网优化调度方法
CN116307437A (zh) 基于碳排放权兑换机制的虚拟电厂优化调度方法及系统
CN115293803A (zh) 电力现货市场下基于电价预测的独立储能运营系统及方法
CN113872226A (zh) 规模化储能参与辅助服务日前增补空间优化方法
CN116720879B (zh) 一种基于双层博弈模型的园区综合能源系统能源定价方法
CN110350537B (zh) 基于可再生能源整合的微电网优化调度方法
CN116454944A (zh) 一种基于随机生产模拟的储能装置优化配置方法及系统
CN117610848A (zh) 一种考虑预测偏差的综合能源系统优化调度方法及其应用