Electro-hydrogen system capacity configuration method considering source charge uncertainty

Abstract

The invention discloses an electro-hydrogen system capacity configuration method considering source load uncertainty, and the method builds a double-layer capacity configuration model, an upper-layer optimization model pays attention to the long-term planning and resource configuration of a system, combines the historical data of renewable energy sources, evaluates the performance of the system in different uncertainty scenes, and improves the capacity configuration efficiency. The lower-layer optimization model focuses on feedback and dynamic adjustment of real-time data, optimizes the operation strategy of the system based on the current power generation capacity and hydrogen demand to adapt to instantaneous market change and demand fluctuation, and can effectively cope with source load uncertainty and improve the flexibility and adaptability of the electro-hydrogen system through the double-layer optimization model. In addition, through scientific configuration and dynamic adjustment of the system capacity, the operation cost can be remarkably reduced, the economical efficiency can be improved, and reliable hydrogen energy supply can be provided for users.

Description

Technical Field

[0001] The present invention belongs to the technical field of capacity configuration of an electric hydrogen system, and in particular relates to a capacity configuration method of an electric hydrogen system taking into account source and load uncertainty. Background Art

[0002] In the context of global energy transformation, the electric hydrogen system (or "electric hydrogen-system") as an emerging energy supply model has received increasing attention. The electric hydrogen system uses renewable energy to electrolyze water to produce hydrogen, achieving efficient conversion and storage of electricity and hydrogen energy, and has important environmental and economic value. The system can not only effectively alleviate the intermittent problem of renewable energy, but also provide flexible regulation capabilities for the power system, promoting efficient use of energy and the realization of emission reduction targets.

[0003] However, the design and operation of the electric hydrogen system faces the challenge of source-load uncertainty. This uncertainty is mainly reflected in the volatility of renewable energy generation (such as wind and solar power) and hydrogen demand, which is manifested in factors such as changes in meteorological conditions, fluctuations in market demand, and technical failures, making system performance and economics difficult to predict. Traditional capacity configuration methods are often based on the assumption of load and resource stability, and lack comprehensive consideration of source-load uncertainty. Summary of the invention

[0004] In order to solve the problems existing in the prior art, the present invention proposes a capacity configuration method for an electric hydrogen system taking into account source-load uncertainty.

[0005] The technical solution of the present invention is as follows:

[0006] A method for configuring the capacity of an electric hydrogen system considering source-load uncertainty, wherein the electric hydrogen system comprises a photovoltaic unit, a wind turbine unit, an electrolyzer, a battery and a hydrogen storage tank, and the method comprises:

[0007] Constructing a two-layer capacity configuration model, wherein the two-layer capacity configuration model includes an upper layer optimization model considering system capacity optimization and a lower layer optimization model considering system operation optimization;

[0008] Based on the historical operation data of the electric hydrogen system, a generative adversarial network is used to generate several electric hydrogen system scenarios;

[0009] Based on the clustering theory, the generated electric-hydrogen system scenarios are reduced to obtain the clustered electric-hydrogen system scenarios;

[0010] Based on the clustered electric-hydrogen system scenarios and the electric-hydrogen system capacity configuration model, a capacity configuration model of the electric-hydrogen system is constructed that takes into account source-load uncertainty;

[0011] The electric-hydrogen system capacity configuration model considering source-load uncertainty is solved to obtain the optimal electric-hydrogen system capacity configuration plan.

[0012] Furthermore, the goal of the upper optimization model is to minimize the annual equal investment cost of equipment, annual operation and maintenance cost, and annual wind and solar power abandonment penalty, as shown in the following formula:

[0013] minG=C inv +C opt +C EA

[0014] In the formula, minG is the minimum planning cost, C inv Represents the annual equivalent investment cost of equipment, C opt is the operation and maintenance cost of the electric hydrogen system, C EA Penalties for curtailing wind and solar power in electric hydrogen systems.

[0015] Furthermore, the expression of the annual equivalent investment cost of the equipment is:

[0016]

[0017] j=PV,BS,HT,AWE,PEM

[0018] In the formula, c inv.j is the unit capacity investment cost of the jth equipment; E j is the planned capacity of the jth type of equipment; n is the planned life of the system; r is the equipment discount rate, C inv.j represents the total investment cost of j types of equipment; PV is photovoltaic, BS is electrical energy storage, HT is hydrogen energy storage, AWE is alkaline water electrolysis, and PEM is proton exchange membrane electrolysis.

[0019] Furthermore, the expression of the lower layer optimization model is as follows:

[0020]

[0021] In the formula, λ opt,j represents the unit power operation and maintenance cost coefficient of the jth type of equipment; represents the real-time operating power of the jth device; ΕΑ Indicates the unit cost of wind and solar power abandonment; Indicates the wind and solar power absorbed by the system in real time; represents the actual real-time wind and solar power generation based on wind and solar resources, C opt is the operation and maintenance cost of the electric hydrogen system, j is the equipment index of the electric hydrogen system, T is the total operation time of the electric hydrogen system, C EA Penalties for curtailing wind and solar power in electric hydrogen systems.

[0022] Furthermore, the operating constraints of the double-layer capacity configuration model include power balance constraints of the electric-hydrogen system and equipment operation constraints.

[0023] Furthermore, the power balance constraints of the electric hydrogen system include

[0024]

[0025] In the formula, represents the grid-connected power generation of the wind turbine at time t; represents the grid-connected power generation of the photovoltaic unit at time t; represents the power consumption of the hydrogen production device at time t, wherein the hydrogen production device comprises an alkaline water electrolysis hydrogen production tank and a hydrogen storage tank; Indicates the power load; Indicates the amount of hydrogen produced by alkaline water electrolysis, represents the amount of hydrogen produced by proton exchange membrane electrolysis, Indicates that the hydrogen storage tank is releasing hydrogen. Indicates that the hydrogen storage tank is filled with hydrogen. represents the hydrogen load, Represents the hydrogen production volume of the hydrogen production device at time t.

[0026] Furthermore, the equipment operation constraints include grid-connected power constraints of photovoltaic units and wind turbine units, constraints of hydrogen production devices, and constraints of batteries;

[0027] The grid-connected power generation constraints of the photovoltaic units and wind turbine units are as follows:

[0028]

[0029] Where P wind (t) is the power generated by the wind turbine in the tth period; is the maximum output of the wind turbine in the tth period; P pv (t) is the power generated by the photovoltaic unit in the tth period; is the maximum output of the photovoltaic unit in the t period;

[0030] The constraints of the hydrogen production device include the constraints of electricity-to-hydrogen conversion efficiency, equipment operating capacity and power constraints, ramping constraints within the unit scheduling cycle of the hydrogen production equipment, hydrogen volume state constraints in the hydrogen storage tank, maximum storage volume constraints of the hydrogen storage tank, and the constraints that the hydrogen storage tank must be charged and discharged at different times; wherein, the constraints of electricity-to-hydrogen conversion efficiency are as follows:

[0031]

[0032] In the formula, It represents the hydrogen production volume of the hydrogen production equipment at time t; represents the electric power of the hydrogen production equipment at time t; EH It represents the specific energy consumption of the electrolyzer; Indicates the density of hydrogen under standard conditions;

[0033] The equipment operating capacity and power constraints are as follows:

[0034]

[0035] Where P EH.max , P EH.min They are the maximum and minimum operating powers of different hydrogen production equipment respectively;

[0036] The ramp constraints within the unit scheduling cycle of the hydrogen production equipment are as follows:

[0037]

[0038] In the formula, The maximum adjustment power of different hydrogen production equipment, Represents the electrical power of the hydrogen production equipment at time t-1;

[0039] The hydrogen volume state constraint model in the hydrogen storage tank is:

[0040]

[0041] Where V H2 (t) represents the volume state of hydrogen in the hydrogen storage tank at time t, V H2 (t-1) represents the volume state of hydrogen in the hydrogen storage tank at time t-1; V H2in (t), V H2out (t) respectively represent the hydrogen filling amount and hydrogen discharge amount of the hydrogen storage tank at time t; η H2in , η H2out They represent the hydrogen charging efficiency and hydrogen releasing efficiency respectively;

[0042] The maximum storage volume constraint model of the hydrogen storage tank is:

[0043]

[0044] Where V H2max 、V H2min They represent the upper and lower limits of the volume of the hydrogen storage tank at time t respectively;

[0045] The hydrogen storage tank is charged and discharged to satisfy the constraint that charging and discharging are not simultaneous:

[0046]

[0047] e in (t)+e out (t)≤1

[0048] e in (t)∈{0,1}

[0049] e out (t)∈{0,1}

[0050] Where: They represent the upper and lower limits of the single hydrogen filling volume of the hydrogen storage tank, They represent the upper and lower limits of the single hydrogen release volume of the hydrogen storage tank; e in (t), e out (t) respectively indicate the hydrogen filling and discharging positions of the hydrogen storage tank;

[0051] The battery constraints include battery state of charge constraints, charge and discharge power constraints, and capacity balance constraints of the battery energy storage system; the battery state of charge constraints are as follows:

[0052]

[0053] In the formula, is the battery state of charge at time t, is the state of charge of the battery at time t+1; BS is the self-consumption rate of the battery; is the charging power and discharging power at time t; C BS Install capacity for batteries; is the battery charging efficiency and discharging efficiency; Δt is a scheduling cycle interval;

[0054] The charge and discharge power constraints are as follows:

[0055]

[0056]

[0057] In the formula, Indicates the upper and lower limits of the battery state of charge; is the maximum charging power and discharging power of the battery, They are 0-1 variables for the charging and discharging status of the battery energy storage unit;

[0058] The capacity balance constraints of the battery energy storage system are as follows:

[0059]

[0060] In the formula, They are the state of charge of the battery system at the start and end of the operation model scheduling calculation.

[0061] Furthermore, the generative adversarial network includes a generator network and a discriminator network; the generator network is used to learn the unknown distribution of the collected historical data samples, and it is expected that a generated electric hydrogen system scenario sample can deceive the true or false judgment of the discriminator network;

[0062] The discriminator network is used to distinguish an input electric hydrogen system scene sample.

[0063] Furthermore, the expression of the capacity configuration model of the electric hydrogen system considering the uncertainty of source and load is:

[0064]

[0065] Where sc is the index of the scene in the electric hydrogen scene set SC, ρ sc is the probability of occurrence of the electric hydrogen scenario, x is the configuration scheme of the electric hydrogen system, f(x,λ sc ) is the configuration objective function of the electric-hydrogen system, and λsc represents the uncertainty parameter of the electric-hydrogen system scenario sc.

[0066] Furthermore, the historical operation data of the electric hydrogen system includes historical operation data of photovoltaics, wind turbines, electric loads, and hydrogen loads related to the electric hydrogen system.

[0067] An electronic device comprises a memory and a processor, wherein the memory stores a computer program, and the processor is used to call and run the computer program stored in the memory to execute any of the methods described above.

[0068] A computer-readable storage medium stores a computer program, wherein the computer program implements the steps of any of the above methods when executed by a processor.

[0069] Compared with the prior art, the present invention has the following beneficial effects:

[0070] The present invention proposes a method for configuring the capacity of an electric hydrogen system that takes into account the uncertainty of source and load. The method establishes a double-layer capacity configuration model. The upper optimization model focuses on the long-term planning and resource allocation of the system, and combines the historical data of renewable energy to evaluate the performance of the system under different uncertainty scenarios. The lower optimization model focuses on the feedback and dynamic adjustment of real-time data. Based on the current power generation capacity and hydrogen demand, the system's operating strategy is optimized to adapt to instantaneous market changes and demand fluctuations. Through the double-layer optimization model, the uncertainty of source and load can be effectively dealt with, and the flexibility and adaptability of the electric hydrogen system can be improved. In addition, through the scientific configuration and dynamic adjustment of the system capacity, the operating cost can be significantly reduced, the economy can be improved, and a reliable hydrogen energy supply can be provided to users.

[0071] The method of the present invention is based on the historical operation data of the electric-hydrogen system, uses a generative adversarial network to generate several electric-hydrogen system scenarios, and then reduces the generated electric-hydrogen system scenarios based on clustering theory to obtain clustered electric-hydrogen system scenarios; based on the clustered electric-hydrogen system scenarios and the electric-hydrogen system capacity configuration model, a capacity configuration model of the electric-hydrogen system that takes into account source and load uncertainty is constructed, thereby effectively alleviating the intermittent problem of renewable energy. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 It is a flow chart of a method for configuring the capacity of an electric hydrogen system taking into account source-load uncertainty in an embodiment;

[0073] Figure 2 Schematic diagram of the structure of the electric hydrogen system in the embodiment. DETAILED DESCRIPTION

[0074] The present invention is further explained below in conjunction with the accompanying drawings and specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention. After reading the present invention, various equivalent modifications of the present invention by those skilled in the art all fall within the scope defined by the claims attached to this application.

[0075] Embodiment 1:

[0076] The present invention provides a method for configuring the capacity of an electric hydrogen system taking into account the uncertainty of source and load. The electric hydrogen system includes a photovoltaic unit, a wind turbine unit, an electrolyzer, a battery and a hydrogen storage tank. Figure 1 As shown, the method includes:

[0077] Construct a two-layer capacity configuration model, which includes an upper-layer optimization model that considers system capacity optimization and a lower-layer optimization model that considers system operation optimization;

[0078] Based on the historical operation data of the electric hydrogen system, a generative adversarial network is used to generate several electric hydrogen system scenarios;

[0079] Based on the clustering theory, the generated electric-hydrogen system scenarios are reduced to obtain the clustered electric-hydrogen system scenarios;

[0080] Based on the clustered electric-hydrogen system scenarios and the electric-hydrogen system capacity configuration model, a capacity configuration model for the electric-hydrogen system considering source-load uncertainty is constructed;

[0081] The capacity configuration model of the electric-hydrogen system considering source-load uncertainty is solved to obtain the optimal capacity configuration plan of the electric-hydrogen system.

[0082] Embodiment 2:

[0083] This embodiment is further designed on the basis of the first embodiment in that the goal of the upper optimization model in this embodiment is to minimize the annual equal investment cost of equipment, annual operation and maintenance cost, and annual wind and solar abandonment penalty, as shown in the following formula:

[0084] minG=C inv +C opt +C EA

[0085] In the formula, minG is the minimum planning cost, C inv Represents the annual equivalent investment cost of equipment, C opt is the operation and maintenance cost of the electric hydrogen system, C EA Penalties for curtailing wind and solar power in electric hydrogen systems.

[0086] Embodiment three:

[0087] This embodiment is further designed on the basis of the second embodiment in that the annual equal investment cost of the equipment in this embodiment takes into account the equipment discount rate, and the expression of the annual equal investment cost of the equipment is:

[0088]

[0089] j=PV,BS,HT,AWE,PEM

[0090] In the formula, c inv.j is the unit capacity investment cost of the jth equipment; E j is the planned capacity of the jth type of equipment; n is the planned life of the system; r is the equipment discount rate, C inv.j represents the total investment cost of j types of equipment; PV is photovoltaic, BS is electrical energy storage, HT is hydrogen energy storage, AWE is alkaline water electrolysis, and PEM is proton exchange membrane electrolysis.

[0091] Embodiment 4:

[0092] This embodiment is further designed on the basis of the first embodiment in that the expression of the lower layer optimization model in this embodiment is as follows:

[0093]

[0094] In the formula, λ opt,j represents the unit power operation and maintenance cost coefficient of the jth type of equipment; represents the real-time operating power of the jth device; ΕΑ Indicates the unit cost of wind and solar power abandonment; Indicates the wind and solar power absorbed by the system in real time; represents the actual real-time wind and solar power generation based on wind and solar resources, C opt is the operation and maintenance cost of the electric hydrogen system, j is the equipment index of the electric hydrogen system, T is the total operation time of the electric hydrogen system, C EAPenalties for curtailing wind and solar power in electric hydrogen systems.

[0095] Embodiment five:

[0096] This embodiment is further designed on the basis of the first embodiment in that the operation constraints of the double-layer capacity configuration model in this embodiment include power balance constraints of the electric hydrogen system and equipment operation constraints.

[0097] Embodiment six:

[0098] This embodiment is further designed on the basis of the fifth embodiment in that the power balance constraints of the electric hydrogen system in this embodiment include:

[0099]

[0100] In the formula, represents the grid-connected power generation of the wind turbine at time t; represents the grid-connected power generation of the photovoltaic unit at time t; Indicates the power consumption of the hydrogen production device at time t, which includes an alkaline water electrolysis hydrogen production tank and a hydrogen storage tank; Indicates the power load; Indicates the amount of hydrogen produced by alkaline water electrolysis, represents the amount of hydrogen produced by proton exchange membrane electrolysis, Indicates that the hydrogen storage tank is releasing hydrogen. Indicates that the hydrogen storage tank is filled with hydrogen. represents the hydrogen load, Represents the hydrogen production volume of the hydrogen production device at time t.

[0101] Embodiment seven:

[0102] This embodiment is further designed on the basis of the fifth embodiment in that the equipment operation constraints in this embodiment include grid-connected power constraints of photovoltaic units and wind turbine units, constraints of hydrogen production devices, and constraints of batteries;

[0103] The grid-connected power generation constraints of photovoltaic units and wind turbines are as follows:

[0104]

[0105] Where P wind (t) is the power generated by the wind turbine in the tth period; is the maximum output of the wind turbine in the tth period; P pv (t) is the power generated by the photovoltaic unit in the tth period; is the maximum output of the photovoltaic unit in the t period;

[0106] The constraints of the hydrogen production device include the constraints of electricity-hydrogen conversion efficiency, equipment operating capacity and power constraints, ramp constraints within the unit scheduling cycle of the hydrogen production equipment, hydrogen volume state constraints in the hydrogen storage tank, maximum storage volume constraints of the hydrogen storage tank, and the constraints that the hydrogen storage tank must be charged and discharged at different times. Among them, the constraints of electricity-hydrogen conversion efficiency are as follows:

[0107]

[0108] In the formula, It represents the hydrogen production volume of the hydrogen production equipment at time t; represents the electric power of the hydrogen production equipment at time t; EH It represents the specific energy consumption of the electrolyzer; Indicates the density of hydrogen under standard conditions;

[0109] The equipment operating capacity and power constraints are as follows:

[0110]

[0111] Where P EH.max , P EH.min They are the maximum and minimum operating powers of different hydrogen production equipment respectively;

[0112] The ramp constraints within the unit dispatch cycle of the hydrogen production equipment are as follows:

[0113]

[0114] In the formula, The maximum adjustment power of different hydrogen production equipment, Represents the electrical power of the hydrogen production equipment at time t-1;

[0115] The hydrogen volume state constraint model in the hydrogen storage tank is:

[0116]

[0117] Where V H2 (t) represents the volume state of hydrogen in the hydrogen storage tank at time t, V H2 (t-1) represents the volume state of hydrogen in the hydrogen storage tank at time t-1; V H2in (t), V H2out (t) respectively represent the hydrogen filling amount and hydrogen discharge amount of the hydrogen storage tank at time t; η H2in , η H2out They represent the hydrogen charging efficiency and hydrogen releasing efficiency respectively;

[0118] The maximum storage volume constraint model of the hydrogen storage tank is:

[0119]

[0120] Where VH2max 、V H2min They represent the upper and lower limits of the volume of the hydrogen storage tank at time t respectively;

[0121] The hydrogen storage tank is charged and discharged to meet the constraints of charging and discharging at different times:

[0122]

[0123] e in (t)+e out (t)≤1

[0124] e in (t)∈{0,1}

[0125] e out (t)∈{0,1}

[0126] Where: They represent the upper and lower limits of the single hydrogen filling volume of the hydrogen storage tank, They represent the upper and lower limits of the single hydrogen release volume of the hydrogen storage tank; e in (t), e out (t) respectively indicate the hydrogen filling and discharging positions of the hydrogen storage tank;

[0127] The constraints of the battery include battery state of charge constraints, charge and discharge power constraints, and capacity balance constraints of the battery energy storage system; the battery state of charge constraints are as follows:

[0128]

[0129] In the formula, is the battery state of charge at time t, is the state of charge of the battery at time t+1; BS is the self-consumption rate of the battery; is the charging power and discharging power at time t; C BS Install capacity for batteries; is the battery charging efficiency and discharging efficiency; Δt is a scheduling cycle interval;

[0130] The charge and discharge power constraints are as follows:

[0131]

[0132]

[0133] In the formula, Indicates the upper and lower limits of the battery state of charge; is the maximum charging power and discharging power of the battery, They are 0-1 variables for the charging and discharging status of the battery energy storage unit;

[0134] The capacity balance constraints of the battery energy storage system are as follows:

[0135]

[0136] In the formula, They are the state of charge of the battery system at the start and end of the operation model scheduling calculation.

[0137] Embodiment eight:

[0138] This embodiment is further designed on the basis of the first embodiment in that the generative adversarial network in this embodiment includes a generator network and a discriminator network; the generator network is used to learn the unknown distribution of the collected historical data samples, and it is expected that a generated electric hydrogen system scenario sample can deceive the true or false judgment of the discriminator network;

[0139] The discriminator network is used to distinguish an input electric hydrogen system scenario sample as much as possible.

[0140] Embodiment nine:

[0141] This embodiment is further designed on the basis of the first embodiment in that the expression of the capacity configuration model of the electric hydrogen system considering the uncertainty of source and load is constructed as follows:

[0142]

[0143] Where sc is the index of the scene in the electric hydrogen scene set SC, ρ sc is the probability of occurrence of the electric hydrogen scenario, x is the configuration scheme of the electric hydrogen system, f(x,λ sc ) is the configuration objective function of the electric-hydrogen system, and λsc represents the uncertainty parameter of the electric-hydrogen system scenario sc.

[0144] Embodiment ten:

[0145] This embodiment is further designed on the basis of the first embodiment in that the historical operation data of the electric hydrogen system in this embodiment includes the historical operation data of photovoltaic, wind turbine, electric load and hydrogen load related to the electric hydrogen system.

[0146] Embodiment eleven:

[0147] An electronic device comprises a memory and a processor, wherein the memory stores a computer program, and the processor is used to call and run the computer program stored in the memory to execute any of the methods described above.

[0148] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of any of the above methods are implemented.

[0149] Application examples:

[0150] This example simulates a certain electric hydrogen system and implements the electric hydrogen system capacity configuration method of the present invention. Figure 2 As shown, the electric hydrogen system includes a photovoltaic unit, a wind turbine unit, an electrolyzer, a battery, a hydrogen storage tank and a control system. In this example, based on the method of the present invention, the grid-connected power constraints of the photovoltaic motor and the wind turbine unit are first constructed as follows:

[0151]

[0152] Among them, P wind (t) is the power generated by the wind turbine in the tth period; is the maximum output of the wind turbine in the tth period; P pv (t) is the power generated by the photovoltaic unit in the tth period; is the maximum output of the photovoltaic unit in the tth period.

[0153] The electric hydrogen system converts electrical energy into hydrogen energy through an electrolyzer, and then couples with the hydrogen energy system. The electrolyzer consumes electrical energy to produce hydrogen and oxygen through electrochemical reactions. The specific reaction equation is shown in (2)

[0154]

[0155] Different types of equipment have different hydrogen production costs and hourly power external characteristics, but for the system capacity configuration problem, the hydrogen production equipment model focuses on the configuration cost and hourly power external characteristics. Therefore, a unified electrolyzer mathematical modeling can be performed for the hydrogen production equipment, mainly including the constraints on electricity-to-hydrogen conversion efficiency, equipment operating capacity and power constraints, and ramp constraints within the unit scheduling cycle of the hydrogen production equipment.

[0156] The electric-to-hydrogen conversion efficiency constraint constructed based on the method of the present invention is shown in formula (3):

[0157]

[0158] In the formula, It represents the hydrogen production volume of the hydrogen production equipment at time t; represents the electric power of the hydrogen production equipment at time t; EH It represents the specific energy consumption of the electrolyzer; Indicates the density of hydrogen under standard conditions.

[0159] The equipment operating capacity and power constraints are shown in formula (4):

[0160]

[0161] Where P EH.max , P EH.min They are the maximum and minimum operating power of different hydrogen production equipment respectively.

[0162] The ramp constraint within the unit scheduling cycle of the hydrogen production equipment is shown in formula (5):

[0163]

[0164] In the formula, The maximum adjustment power of different hydrogen production equipment, Represents the electrical power of the hydrogen production equipment at time t-1.

[0165] In the electric hydrogen system, gaseous hydrogen storage is the preferred storage method for long-term and large-scale storage of hydrogen due to its high efficiency and low loss. Hydrogen has high energy density and is suitable for long-term storage, and is complementary to electric energy storage. Considering the long storage time of hydrogen energy storage, if the model is set to be the same as the electric energy storage model, that is, the energy setting at the end of the optimization of hydrogen energy storage a day ago is equal to the energy at the initial time, the model will limit the flexibility of hydrogen energy storage under certain conditions and cannot reflect the daytime energy transfer characteristics of hydrogen energy storage. Therefore, the mathematical model of the hydrogen storage tank is set as follows in the present invention.

[0166] In this example, based on the method of the present invention, the specific steps of constructing the hydrogen volume state constraint in the hydrogen storage tank, the maximum storage volume constraint of the hydrogen storage tank, and the constraint that the hydrogen storage tank can be charged and discharged at different times are as follows:

[0167] The volume state of hydrogen in the hydrogen storage tank at time t can be expressed as:

[0168]

[0169] Among them, V H2 (t) represents the volume state of hydrogen in the hydrogen storage tank at time t, V H2 (t-1) represents the volume state of hydrogen in the hydrogen storage tank at time t-1; V H2in (t), V H2out (t) respectively represent the hydrogen filling amount and hydrogen discharge amount of the hydrogen storage tank at time t; η H2in , η H2out They represent the hydrogen charging efficiency and hydrogen decomposition efficiency respectively.

[0170] The volume of hydrogen in the hydrogen storage tank at each moment needs to be limited to the maximum storage volume of the hydrogen storage tank

[0171]

[0172] Among them, V H2max 、V H2min They respectively represent the upper and lower limits of the hydrogen storage tank volume at time t.

[0173] The charging and discharging of hydrogen storage tanks must meet the constraint that charging and discharging cannot be performed simultaneously:

[0174]

[0175] e in (t)+e out (t)≤1 (10)

[0176] e in (t)∈{0,1}(11)

[0177] e out (t)∈{0,1}(12)

[0178] in, They represent the upper and lower limits of the single hydrogen filling volume of the hydrogen storage tank, They represent the upper and lower limits of the single hydrogen release volume of the hydrogen storage tank; e in (t), e out (t) respectively represent the hydrogen storage tank charging and discharging identification positions.

[0179] The battery-related constraint model constructed in this example based on the method of the present invention is as follows:

[0180] The battery constraints include battery state of charge (SOC) constraints, charge and discharge power constraints, and capacity balance constraints of the battery energy storage system.

[0181] Equations (13) and (14) represent the battery state of charge (SOC) constraints during the battery charging and discharging process:

[0182]

[0183] In the formula, is the battery state of charge at time t, is the state of charge of the battery at time t+1; BS is the self-consumption rate of the battery; is the charging power and discharging power at time t; C BS Install capacity for batteries; is the battery charging efficiency and discharging efficiency; Δt is a scheduling cycle interval.

[0184] Formula (15-18) represents the SOC and charge and discharge power constraints

[0185]

[0186] Where: Indicates the upper and lower limits of the battery state of charge; is the maximum charging power and discharging power of the battery, They are 0-1 variables for the charging and discharging status of the battery energy storage unit.

[0187] Formula (19) represents the capacity balance of the battery energy storage system during the entire dispatch period:

[0188]

[0189] In the formula, They are the state of charge of the battery system at the start and end of the operation model scheduling calculation.

[0190] The lower layer optimization model constructed based on the method of the present invention in this example is as follows:

[0191]

[0192] In the formula, λ opt,j represents the unit power operation and maintenance cost coefficient of the jth type of equipment; represents the real-time operating power of the jth device; ΕΑ Indicates the unit cost of wind and solar power abandonment; Indicates the wind and solar power absorbed by the system in real time; represents the actual real-time wind and solar power generation based on wind and solar resources, C opt is the operation and maintenance cost of the electric hydrogen system, j is the equipment index of the electric hydrogen system, T is the total operation time of the electric hydrogen system, C EA Penalties for curtailing wind and solar power in electric hydrogen systems.

[0193] Operation constraints are oriented to system production simulation. First, the system power balance equation and hydrogen production balance equation are established as shown in equations (22)-(23):

[0194]

[0195]

[0196] Where: represents the grid-connected power generation of wind turbines at time t; represents the grid-connected power generation of the photovoltaic unit at time t; Indicates the power consumption of the hydrogen production device at time t; Indicates the power load; Indicates the amount of hydrogen produced by alkaline water electrolysis, represents the amount of hydrogen produced by proton exchange membrane electrolysis, Indicates that the hydrogen storage tank is releasing hydrogen. Indicates that the hydrogen storage tank is filled with hydrogen. represents the hydrogen load, It represents the hydrogen production volume of the hydrogen production device at time t;

[0197] In this example, the objective function of the upper optimization model is to minimize the annual equivalent cost of the entire life cycle, including the annual equivalent investment cost of equipment, the annual equivalent operation and maintenance cost, and the penalty for wind and solar power abandonment, as shown in formula (24):

[0198] minF=C inv +C opt +C EA (twenty four)

[0199] Among them, C inv represents the equipment investment cost, and minF is the minimum cost.

[0200] The equipment annual equivalent investment cost takes into account the equipment discount rate, which can be calculated by formula (25):

[0201]

[0202] j=PV,BS,HT,AWE,PEM (27)

[0203] Where: c inv.j is the unit capacity investment cost of the jth equipment; E j is the planned capacity of the jth type of equipment; n is the planned life of the system; r is the equipment discount rate, C inv.j represents the total investment cost of j types of equipment; in formula (27), PV is photovoltaic, BS is electrical energy storage, HT is hydrogen energy storage, AWE is alkaline water electrolysis, and PEM is proton exchange membrane electrolysis.

[0204] This example collects historical operating data of the electric hydrogen system based on the method of the present invention, including historical operating data of photovoltaics, wind turbines, electric loads, and hydrogen loads.

[0205] In this example, a generative adversarial network is used to generate the electric hydrogen system scenario (also called "scenario"). The generative adversarial network generally includes an adversarial framework composed of two interconnected neural networks. These two networks are generally composed of deep neural networks: one network is a generator network, which is used to learn the unknown distribution of data samples and expect the generated samples to deceive the true and false judgment of the discriminator network; the other network is a discriminator network, which is used to distinguish the real input samples and the generated samples as much as possible. The discriminator network feeds back the judgment results of the input samples to the generator network, so that the generator network continuously adjusts its own parameters to generate more realistic samples. The discriminator network adjusts its own parameters according to the continuously generated samples to better discover the differences in the input samples. In this way, the generator network continuously improves the quality of the generated samples, while the discriminator network continuously improves its judgment ability. As the training continues to converge, the generator network finally captures the potential distribution of the real data. At this time, whether for the generated samples or the real samples, the probability of the discriminator network's judgment is close to 0.5, the training reaches the optimal state, and the discriminator network cannot judge the true or false.

[0206] The distribution law of historical data x is recorded as P r, and define a noise variable z. The generator samples through the fully connected layer network and the deconvolution layer network, maps the noise variable z to a generated data space G(z), and records its data distribution as P G .

[0207] The zero-sum game between the two networks is denoted as V(D,G). The optimization of the loss function needs to be based on the update of the other network, while only updating the parameters of the own network. The two networks are continuously trained alternately and their parameters are continuously updated until the adversarial network reaches Nash equilibrium.

[0208] The two adversarial networks update their own parameters and use weights to remember the learning experience during the training process. Ideally, when the two deep neural network training finally reaches Nash equilibrium, the discriminator network cannot distinguish the input samples, and the probability judgment of the input samples is 0.5, and the model reaches the optimal state.

[0209] Assume that for a certain power generation site, actual observation data P in the past is provided at time t hist At the same time, given the historical observation condition p hist Below, point forecasts are provided for each lead time point Based on historical information and point prediction data, a set of N scenarios S = {s1,…,s N} can represent possible future behaviors around point predictions.

[0210] For a typical wind or photovoltaic power generation site, the collected historical observations and the point forecasts at each lead time point can be expressed as follows:

[0211]

[0212] Where: h represents the time range of historical data, and k represents the forecast range.

[0213] Assume that a GAN model is trained based on the observed data set, without considering point prediction In the case of , given the input noise z, G(z) will output any possible power generation realization. Therefore, it is necessary to constrain the possible realizations G(z) so that the output trajectory meets the predicted requirements.

[0214] Assuming that the time range of the historical data part of the sample is represented by l, the following two conditions need to be met for generating data G(z):

[0215] (1) From time point tl to t, the output of the corresponding G(z) should be as close as possible to the historical data P hist ;

[0216] (2) From time point t+1 to t+k, the output part of the corresponding G(z) should be as realistic as possible and close to the center point prediction.

[0217] Split a vector into two parts, given v∈R l+k+1 , so the first l+1 dimension and the last k dimension of the mapping vector v can be expressed as

[0218]

[0219] Among them, P hist (f(v)) represents the historical data set, P pred (f(v)) represents the prediction dataset.

[0220] In order to ensure that the first part of G(z) is consistent with the historical data P hist =[p t-l …p t ] to be consistent, using the following objective function:

[0221] min||P hist (G(z))-P hist ||2(31)

[0222] Among them, P hist (G(z)) represents the adversarial historical dataset.

[0223] According to the basic principle of generative adversarial network, the larger the output of the discriminator, the more realistic the sample. In order to generate a realistic wind power trajectory, a loss term -D(G(z)) is added to the objective function. Since wind power generation has the inherent uncertainty, an upper bound is used and a lower bound Describes a prediction space. The prediction interval can be controlled by a prediction confidence parameter η:

[0224]

[0225] Combining the objectives and constraints, given the historical observation condition P hist , point prediction As well as the generator and the discriminator, the scene prediction problem can be expressed as a constrained optimization problem

[0226]

[0227] in, represents the loss coefficient, and D(G(z)) represents the loss term of the objective function.

[0228] The generator uses three deconvolution networks to upsample the input noise z to generate wind power scenarios. The discriminator has the opposite structure to the generator and downsamples the historical data and samples generated by the generator through three convolutional layers.

[0229] In this example, the K-medoids clustering method is used to reduce the scenarios of the electric-hydrogen system. A large number of renewable energy scenarios can be generated using a generative adversarial network. However, since large-scale scenarios will reduce the efficiency of solving the optimal configuration of the electric-hydrogen system capacity, the practicality is limited. It is necessary to use a clustering algorithm to reduce the number of new energy scenarios, thereby improving the overall quality of new energy scenarios and improving the efficiency of random optimization of the electric-hydrogen system fusion configuration. Assume that any two scenarios u in the scenario set S i ,u j The distance between i ,u j ), a set J consisting of K scenes r Replacing the original scene set S is scene reduction. The optimal set J r It should be the set whose probability information is closest to the original scene set S, that is, it satisfies the following optimization problem:

[0230]

[0231] Where: d(u i ,u j ) is the scene u i With scene u j The distance between i For scene u i probability.

[0232] Therefore, scene reduction seeks the set J r The essence of K-medoids is similar to the cluster center set in the clustering algorithm. The K-medoids algorithm is not sensitive to changes in the problem scale and can be introduced into a parallel architecture. Need to be with each scene j Compare. If set J r The larger the scale, the more computational effort will be spent on traversal. To solve this problem, we can introduce a reference scene u 0 , according to the Minkowski inequality,

[0233] |d(u i ,u 0 )-d(u j ,u 0 )|≤d(u i ,u j ) (35)

[0234] Update the set C in the K-medoids algorithm k Center When C k The weighted distance sum of each scene and other scenes in the , the smallest scene is the new center

[0235]

[0236] In this example, a capacity configuration model of the electric-hydrogen system considering the uncertainty of source and load is established, as shown in formula (37):

[0237]

[0238] Where sc is the index of the scene in the scene set SC, ρ sc is the probability of the scenario occurring, x is the configuration scheme of the electric hydrogen system, f(x,λ sc ) is the objective function, and λsc represents the uncertainty parameter of the scene sc.

[0239] In this example, the particle swarm algorithm is used to solve the optimization model, including: initializing algorithm parameters, setting the upper and lower limits of particles in the particle swarm algorithm, and randomly generating a group of particles as initial particles; setting the optimization target of the double-layer optimization model as the fitness function of the particle swarm algorithm, and calculating the fitness value of the initial particle swarm; comparing the fitness value of each particle with the fitness value of the individual optimal solution, and retaining the optimal solution; then updating the speed and position of the particles according to the principle of the particle swarm algorithm; and finally obtaining the optimal configuration scheme of the electric hydrogen system through continuous iteration.

[0240] The above description is only a specific implementation of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by a technician familiar with the technical field within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention.

Claims

1. A method for configuring the capacity of an electric hydrogen system considering source-load uncertainty, wherein the electric hydrogen system comprises a photovoltaic unit, a wind turbine unit, an electrolyzer, a battery and a hydrogen storage tank, characterized in that: The method comprises: Constructing a two-layer capacity configuration model, wherein the two-layer capacity configuration model includes an upper layer optimization model considering system capacity optimization and a lower layer optimization model considering system operation optimization; Based on the historical operation data of the electric hydrogen system, a generative adversarial network is used to generate several electric hydrogen system scenarios; Based on the clustering theory, the generated electric-hydrogen system scenarios are reduced to obtain the clustered electric-hydrogen system scenarios; Based on the clustered electric-hydrogen system scenarios and the electric-hydrogen system capacity configuration model, a capacity configuration model of the electric-hydrogen system is constructed that takes into account source-load uncertainty; The electric-hydrogen system capacity configuration model considering source-load uncertainty is solved to obtain the optimal electric-hydrogen system capacity configuration plan.

2. The method for configuring the capacity of an electric hydrogen system considering source-load uncertainty according to claim 1 is characterized in that: The goal of the upper optimization model is to minimize the annual equal investment cost of equipment, annual operation and maintenance cost, and annual wind and solar abandonment penalty, as shown in the following formula minG=C inv +C opt +C EA In the formula, min G is the minimum planning cost, C inv Represents the annual equivalent investment cost of equipment, C opt is the operation and maintenance cost of the electric hydrogen system, C EA Penalties for curtailing wind and solar power in electric hydrogen systems.

3. The method for configuring the capacity of an electric hydrogen system considering source-load uncertainty according to claim 2 is characterized in that: The expression of the annual equivalent investment cost of the equipment is: j=PV,BS,HT,AWE,PEM In the formula, c inv.j is the unit capacity investment cost of the jth equipment; E j is the planned capacity of the jth type of equipment; n is the planned life of the system; r is the equipment discount rate, C inv.j represents the total investment cost of j types of equipment; PV is photovoltaic, BS is electrical energy storage, HT is hydrogen energy storage, AWE is alkaline water electrolysis, and PEM is proton exchange membrane electrolysis.

4. The method for configuring the capacity of an electric hydrogen system considering source-load uncertainty according to claim 1, characterized in that: The expression of the lower layer optimization model is as follows: In the formula, λ opt,j represents the unit power operation and maintenance cost coefficient of the jth type of equipment; represents the real-time operating power of the jth device; ΕΑ Indicates the unit cost of wind and solar power abandonment; Indicates the wind and solar power absorbed by the system in real time; represents the actual real-time wind and solar power generation based on wind and solar resources, C opt is the operation and maintenance cost of the electric hydrogen system, j is the equipment index of the electric hydrogen system, T is the total operation time of the electric hydrogen system, C EA Penalties for curtailing wind and solar power in electric hydrogen systems.

5. The method for configuring the capacity of an electric hydrogen system considering source-load uncertainty according to claim 1, characterized in that: The operating constraints of the double-layer capacity configuration model include power balance constraints of the electric-hydrogen system and equipment operation constraints.

6. The method for configuring the capacity of an electric hydrogen system considering source-load uncertainty according to claim 5 is characterized in that: The power balance constraints of the electric hydrogen system include In the formula, represents the grid-connected power generation of the wind turbine at time t; represents the grid-connected power generation of the photovoltaic unit at time t; represents the power consumption of the hydrogen production device at time t, wherein the hydrogen production device comprises an alkaline water electrolysis hydrogen production tank and a hydrogen storage tank; Indicates the power load; Indicates the amount of hydrogen produced by alkaline water electrolysis, represents the amount of hydrogen produced by proton exchange membrane electrolysis, Indicates that the hydrogen storage tank is releasing hydrogen. Indicates that the hydrogen storage tank is filled with hydrogen. represents the hydrogen load, Represents the hydrogen production volume of the hydrogen production device at time t.

7. The method for configuring the capacity of an electric hydrogen system considering source-load uncertainty according to claim 5, characterized in that: The equipment operation constraints include grid-connected power constraints of photovoltaic units and wind turbine units, constraints of hydrogen production devices, and constraints of batteries; The grid-connected power generation constraints of the photovoltaic units and wind turbine units are as follows: Where P wind (t) is the power generated by the wind turbine in the tth period; is the maximum output of the wind turbine in the tth period; P pv (t) is the power generated by the photovoltaic unit in the tth period; is the maximum output of the photovoltaic unit in the t period; The constraints of the hydrogen production device include the constraints of electricity-hydrogen conversion efficiency, equipment operating capacity and power constraints, ramping constraints within the unit scheduling cycle of the hydrogen production equipment, hydrogen volume state constraints in the hydrogen storage tank, maximum storage volume constraints of the hydrogen storage tank, and the constraints that the hydrogen storage tank must be charged and discharged at different times; wherein, the constraints of electricity-hydrogen conversion efficiency are as follows: In the formula, It represents the hydrogen production volume of the hydrogen production equipment at time t; represents the electric power of the hydrogen production equipment at time t; EH It represents the specific energy consumption of the electrolyzer; Indicates the density of hydrogen under standard conditions; The equipment operating capacity and power constraints are as follows: Where P EH.max , P EH.min They are the maximum and minimum operating powers of different hydrogen production equipment respectively; The ramp constraints within the unit scheduling cycle of the hydrogen production equipment are as follows: In the formula, The maximum adjustment power of different hydrogen production equipment, Represents the electrical power of the hydrogen production equipment at time t-1; The hydrogen volume state constraint model in the hydrogen storage tank is: Where V H2 (t) represents the volume state of hydrogen in the hydrogen storage tank at time t, V H2 (t-1) represents the volume state of hydrogen in the hydrogen storage tank at time t-1; V H2in (t), V H2out (t) respectively represent the hydrogen filling amount and hydrogen discharge amount of the hydrogen storage tank at time t; η H2in , η H2out They represent the hydrogen charging efficiency and hydrogen decomposition efficiency respectively; The maximum storage volume constraint model of the hydrogen storage tank is: Where V H2max 、V H2min They represent the upper and lower limits of the volume of the hydrogen storage tank at time t respectively; The hydrogen storage tank is charged and discharged to satisfy the constraint that charging and discharging are not simultaneous: e in (t)+e out (t)≤1 e in (t)∈{0,1} e out (t)∈{0,1} Where: They represent the upper and lower limits of the single hydrogen filling volume of the hydrogen storage tank, They represent the upper and lower limits of the single hydrogen release volume of the hydrogen storage tank; e in (t), e out (t) respectively indicate the hydrogen storage tank charging and discharging identification positions; The battery constraints include battery state of charge constraints, charge and discharge power constraints, and capacity balance constraints of the battery energy storage system; the battery state of charge constraints are as follows: In the formula, is the battery state of charge at time t, is the state of charge of the battery at time t+1; BS is the self-consumption rate of the battery; is the charging power and discharging power at time t; C BS Install capacity for batteries; is the battery charging efficiency and discharging efficiency; Δt is a scheduling cycle interval; The charge and discharge power constraints are as follows: In the formula, Indicates the upper and lower limits of the battery state of charge; is the maximum charging power and discharging power of the battery, They are 0-1 variables for the charging and discharging status of the battery energy storage unit; The capacity balance constraints of the battery energy storage system are as follows: In the formula, They are the state of charge of the battery system at the start and end of the operation model scheduling calculation.

8. The method for configuring the capacity of an electric hydrogen system considering source-load uncertainty according to claim 1, characterized in that: The generative adversarial network includes a generator network and a discriminator network; the generator network is used to learn the unknown distribution of the collected historical data samples, and it is expected that a generated electric hydrogen system scenario sample can deceive the true or false judgment of the discriminator network; The discriminator network is used to distinguish an input electric hydrogen system scene sample.

9. The method for configuring the capacity of an electric hydrogen system considering source-load uncertainty according to claim 1, characterized in that: The expression of the capacity configuration model of the electric hydrogen system considering the uncertainty of source and load is: Where sc is the index of the scene in the electric hydrogen scene set SC, ρ sc is the probability of occurrence of the electric hydrogen scenario, x is the configuration scheme of the electric hydrogen system, f(x,λ sc ) is the configuration objective function of the electric-hydrogen system, and λsc represents the uncertainty parameter of the electric-hydrogen system scenario sc.

10. The method for configuring the capacity of an electric hydrogen system considering source-load uncertainty according to claim 1, characterized in that: The historical operation data of the electric hydrogen system includes historical operation data of photovoltaic, wind turbine, electric load and hydrogen load related to the electric hydrogen system.

Images