Capacity configuration method of green electricity hydrogen production comprehensive energy system under supply and demand uncertainty

Abstract

The application relates to the field of comprehensive energy system planning, and specifically discloses a green electricity hydrogen production comprehensive energy system capacity configuration method under supply and demand uncertainty, which comprises the following steps: obtaining meteorological data, load data and equipment parameters; constructing a system architecture containing wind and light power generation, light heat, electrolytic tank, hydrogen gas turbine and multi-element energy storage, and establishing a mathematical model; generating a typical scenario set based on historical data and constructing a discrete probability fuzzy set based on norm constraint; constructing an upper capacity planning model with economy as the target and a lower operation scheduling optimization model with minimum energy shortage as the target, to form a two-stage distribution robust optimization model; introducing a conditional risk value constraint to control the extreme energy shortage risk; and iteratively solving by using a column and constraint generation algorithm to output an optimal capacity configuration scheme. The application can effectively cope with the uncertainty of both supply and demand, collaboratively optimize economy and reliability, and is suitable for the planning and construction of a park-level comprehensive energy system.

Description

Technical Field

[0001] This application relates to the field of integrated energy system planning, and in particular to the capacity configuration method of integrated energy systems for green electricity to hydrogen production under uncertain supply and demand conditions. Background Technology

[0002] With the large-scale grid connection of renewable energy sources such as wind and solar power, their inherent intermittency and volatility pose severe challenges to the planning and operation of power systems. Hydrogen energy, especially green hydrogen produced by electrolysis of renewable energy, has unique advantages in promoting the consumption of new energy sources and achieving decarbonization of end-use energy. However, the uncertainty on both the supply and demand sides caused by the high proportion of renewable energy integration puts the capacity configuration of green hydrogen production systems under dual pressure of economic efficiency and reliability.

[0003] Existing technologies are largely limited to a unidirectional "electricity-hydrogen" coupling model in terms of system architecture, with insufficient attention paid to solar thermal participation and the cascade utilization of thermal energy. Regarding uncertainty handling, traditional stochastic programming relies on precise probability distribution assumptions, while robust optimization is overly conservative, making it difficult to achieve a balance between robustness and economy. In terms of optimization objectives, most studies focus solely on economic efficiency or use rigid constraints to address reliability, lacking an intrinsic trade-off between the two. In terms of decision-making, most studies presuppose equipment types or optimize only a single dimension, failing to meet the modular configuration requirements of engineering practice. Therefore, there is an urgent need for a comprehensive energy system capacity configuration method that can effectively address uncertainties on both the supply and demand sides, achieve deep multi-energy coupling and cascade utilization, and synergistically optimize both economy and reliability. Summary of the Invention

[0004] To address this, this application provides a method for configuring the capacity of a green electricity-to-hydrogen integrated energy system under uncertain supply and demand conditions, in an attempt to solve or at least alleviate the problems mentioned above.

[0005] According to one aspect of this application, a method for configuring the capacity of a green electricity-to-hydrogen integrated energy system under supply and demand uncertainty is provided, including:

[0006] S1. Obtain meteorological data, load demand data, and technical and economic parameters of each candidate device for the target area;

[0007] S2. Based on meteorological data and load demand data, construct the physical architecture of the green electricity-to-hydrogen integrated energy system, and establish mathematical models of each device based on the physical architecture;

[0008] S3. Based on the historical statistical characteristics of meteorological data and load demand data, determine the uncertain parameter vector, generate a typical scenario set and its initial probability distribution based on the uncertain parameter vector, and construct a discrete probability fuzzy set based on norm constraints based on the typical scenario set.

[0009] S4. Based on the mathematical models of each device, the technical and economic parameters of each candidate device, the typical scenario set, and the discrete probability fuzzy set, construct a two-stage sub-Bruker capacity configuration optimization model.

[0010] S5. Based on the discrete probability fuzzy set, conditional risk value constraints are introduced into the two-stage sub-Bruker capacity allocation optimization model to obtain a two-stage sub-Bruker capacity allocation optimization model with conditional risk value constraints.

[0011] S6. Use the column and constraint generation algorithm to solve the two-stage sub-Bruker capacity allocation optimization model with conditional risk value constraints. Iterate until the convergence condition is met and output the optimal capacity allocation scheme.

[0012] Optionally, in the method according to this application, in S2, the equipment in the physical architecture includes power supply equipment, conversion equipment, energy storage equipment, and load equipment; the power supply equipment includes wind power generation equipment, photovoltaic power generation equipment, and solar thermal equipment; the conversion equipment includes electrolyzers and hydrogen gas turbines; the energy storage equipment includes batteries, hydrogen storage tanks, and thermal storage tanks; the load equipment includes electrical load equipment, thermal load equipment, and hydrogen load equipment; the mathematical model of each device includes the operating constraint equations of each device, as well as the power balance equations, thermal power balance equations, and hydrogen mass balance equations of the green electricity hydrogen production integrated energy system established based on the operating constraint equations of each device.

[0013] Optionally, in the method according to this application, step S3, generating a typical scene set and its initial probability distribution based on the uncertain parameter vector, specifically includes:

[0014] Fitting the cumulative distribution function of each random variable based on historical data of uncertain parameter vectors;

[0015] Latin hypercube sampling is used to generate an initial scene set: the range of values ​​of each random variable is divided into N intervals with equal probability, representative points are randomly selected in each interval, and sampled values ​​are obtained through inverse transformation sampling. The N sampled values ​​are randomly arranged to generate N initial scenes, and the initial probability of each initial scene is 1 / N.

[0016] The initial scene set is reduced by using the synchronous back-substitution elimination method. In each iteration, the probability distance loss between each scene and its nearest scene is calculated. The scene with the smallest probability distance loss is deleted, and the probability of the deleted scene is added to its nearest scene until the number of remaining scenes reaches the preset target value, thus obtaining the typical scene set and its initial probability distribution.

[0017] Optionally, in the method according to this application, during the process of reducing the initial scenario set using the synchronous back-substitution elimination method, scenarios in the initial scenario set where the renewable energy output is lower than a first preset threshold or the load demand is higher than a second preset threshold are identified in advance as extreme scenarios, and extreme scenarios are forcibly retained during the reduction process.

[0018] Optionally, in the method according to this application, in S3, the discrete probability fuzzy set based on norm constraints is:

[0019] In a typical scenario set, the set of probability distributions that satisfy the non-negativity and normalization conditions, whose 1-norm distance from the initial probability distribution does not exceed the first radius and whose ∞-norm distance does not exceed the second radius; the first radius is used to control the total probability offset of all scenarios in the typical scenario set, and the second radius is used to control the maximum probability offset of a single scenario in the typical scenario set.

[0020] Optionally, in the method according to this application, in S4, the two-stage distributed capacity configuration optimization model includes an upper-level capacity planning model and a lower-level operation scheduling optimization model;

[0021] The upper-level capacity planning model takes the mathematical models of each device and the technical and economic parameters of each candidate device as inputs, aims to minimize the sum of investment cost, maintenance cost and expected operating cost over the entire life cycle of the system, and uses equipment selection variables, installation quantity variables and single unit capacity variables as decision variables to generate a capacity configuration scheme.

[0022] The lower-level operation scheduling optimization model takes the capacity configuration scheme, typical scenario set and mathematical model of each device as input, takes minimizing the energy shortage under each typical scenario as the objective, and takes the hourly operating power of each device as the decision variable to solve for the minimum energy shortage and minimum operating cost under each typical scenario. The minimum operating cost is then fed back to the upper-level capacity planning model to calculate the expected operating cost.

[0023] The energy shortage is the sum of the equivalent energy calculated from the power shortage of electrical load, the power shortage of thermal load, and the rate of power shortage of hydrogen load within the dispatch cycle.

[0024] Optionally, in the method according to this application, in S5, the conditional value-at-risk constraint is expressed as:

[0025] Under all possible probability distributions within the discrete probability fuzzy set, the conditional risk value of the energy shortage output by the lower-level operation scheduling optimization model does not exceed a preset threshold; the conditional risk value is used to measure the expected level of the energy shortage at the tail of the probability distribution.

[0026] Optionally, in the method according to this application, in S6, a column and constraint generation algorithm is used to solve the two-stage sub-Bruker capacity allocation optimization model with conditional risk value constraints, specifically including:

[0027] S61. Set the worst-case scenario set to an empty set, set the upper bound to positive infinity, and set the lower bound to negative infinity;

[0028] S62. Under the current worst-case scenario set, solve the upper-level capacity planning model, and use the deterministic equivalent form of the conditional risk value constraint under the current worst-case scenario set as an additional constraint to obtain the current capacity configuration scheme and update the lower bound.

[0029] S63. Under the current capacity configuration scheme, solve the lower-level operation scheduling optimization model for each typical scenario to obtain the minimum energy shortage for each typical scenario; based on the minimum energy shortage for each typical scenario, search for the probability distribution that maximizes the conditional risk value in the discrete probability fuzzy set to obtain the worst probability distribution and the maximum conditional risk value; update the upper bound according to the expected operating cost under the current capacity configuration scheme and the worst probability distribution.

[0030] S64. If the difference between the upper and lower bounds is less than the convergence accuracy, then the current capacity configuration scheme is the optimal solution. Stop the iteration and output the optimal capacity configuration scheme. If the difference between the upper and lower bounds is greater than or equal to the convergence accuracy, then select the scenario with the largest energy shortage from the typical scenario set and add it to the worst scenario set. Return to S62 and execute again.

[0031] According to another aspect of this application, a computing device is provided, comprising: at least one processor; and a memory storing program instructions, wherein the program instructions are configured to be executed by the at least one processor, the program instructions including instructions for executing the deep reinforcement learning-based AI-Agent smart bidding prediction method for the electricity market according to this application.

[0032] According to another aspect of this application, a readable storage medium storing program instructions is provided, which, when read and executed by a computing device, causes the computing device to execute the deep reinforcement learning-based AI-Agent intelligent bidding prediction method for the electricity market according to this application.

[0033] In summary, this application constructs an integrated energy system architecture encompassing solar thermal equipment, an electrolyzer, a hydrogen gas turbine, and multi-source energy storage. On the energy supply side, solar thermal energy is introduced as a stable heat source. On the conversion side, a bidirectional coupling channel is established between hydrogen production via the electrolyzer and cogeneration via the hydrogen gas turbine. On the heat flow side, a closed loop of multi-source heat energy synergistic supply and cascaded utilization is formed. This overcomes the limitations of traditional unidirectional "electricity-hydrogen" coupling, achieving deep synergy and efficient utilization of electricity, heat, and hydrogen energy flows, significantly improving the overall energy efficiency of the system. Furthermore, this application employs a norm-constrained discrete probability fuzzy set to describe the uncertainties on both the supply and demand sides. By flexibly adjusting the fuzzy set radius, it utilizes historical data while avoiding dependence on precise probability distributions, effectively overcoming the inherent defects of traditional stochastic programming and robust optimization, and achieving robustness... The application achieves a dynamic balance between robustness and economy. It constructs an upper-level capacity planning model with economy as the objective and a lower-level operation scheduling optimization model with minimizing energy shortage as the objective, forming a two-level hierarchical collaborative optimization objective system. Conditional risk value is embedded as a reliability metric into the sub-Bruker framework, elevating reliability from a rigid constraint to an optimizable objective, realizing an endogenous trade-off between economy and reliability, and effectively controlling the risk of energy shortage in extreme scenarios. The application incorporates equipment selection, installation quantity, and single-unit capacity into a unified optimization framework, realizing three-dimensional joint optimization. This breaks through the limitations of existing research that presupposes equipment types or optimizes only a single dimension, providing a more accurate quantitative tool for modular configuration decisions in engineering practice and improving the fit between planning schemes and actual engineering. Attached Figure Description

[0034] To achieve the foregoing and related objectives, certain illustrative aspects are described herein in conjunction with the following description and accompanying drawings. These aspects indicate various ways in which the principles disclosed herein may be practiced, and all aspects and their equivalents are intended to fall within the scope of the claimed subject matter. The foregoing and other objectives, features, and advantages of this disclosure will become more apparent from the following detailed description, taken in conjunction with the accompanying drawings. Throughout this disclosure, the same reference numerals generally refer to the same parts or elements.

[0035] Figure 1 A flowchart illustrating a method for configuring the capacity of a green electricity-to-hydrogen integrated energy system under supply and demand uncertainty, according to an embodiment of this application, is shown.

[0036] Figure 2 A diagram illustrating a framework for constructing a fuzzy set of a sub-Brussels bar according to an embodiment of this application is shown.

[0037] Figure 3 A flowchart illustrating the column and constraint generation algorithm solution according to an embodiment of this application is shown. Detailed Implementation

[0038] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0039] With the large-scale grid connection of renewable energy sources such as wind and solar power, their inherent intermittency and volatility pose severe challenges to the planning and operation of power systems. Hydrogen energy, as a clean and efficient secondary energy carrier, especially green hydrogen produced by electrolysis of renewable energy, has unique advantages in promoting the consumption of new energy sources and achieving decarbonization of end-use energy. The green electricity-hydrogen integrated energy system organically integrates wind and solar power generation, solar thermal heating, hydrogen electrolysis, hydrogen gas turbine cogeneration, diversified energy storage, and heat recovery, achieving cascaded utilization and flexible dispatch of energy through deep coupling of electricity, hydrogen, and heat.

[0040] However, existing technologies still have the following shortcomings:

[0041] Insufficiency at the system architecture level. Existing research is mostly limited to the "electricity-hydrogen" unidirectional coupling mode, with insufficient consideration for the participation of photothermal energy and the cascade utilization of thermal energy. There is a lack of unified modeling for the synergistic and cascade utilization of multi-source thermal energy, resulting in a large amount of medium and low temperature waste heat resources not being effectively recovered and utilized, and the overall energy efficiency of the system being low.

[0042] Insufficiency in uncertainty modeling. The supply and demand uncertainties caused by the high proportion of renewable energy integration place dual pressures on capacity allocation, posing challenges to both economic efficiency and reliability. Traditional stochastic programming methods rely on precise probability distribution assumptions, which are difficult to obtain accurately in practice. Traditional robust optimization methods use box-type uncertainty sets to describe parameter fluctuation ranges and seek the optimal solution in the worst case, often leading to overly conservative allocation schemes. Existing research rarely applies the joint constraints of 1-norm and ∞-norm to capacity allocation problems, making it difficult to flexibly balance robustness and economy.

[0043] There are shortcomings at the target system level. Existing research mostly focuses on single economic efficiency or uses reliability as a rigid constraint. A single economic objective is difficult to characterize the operational risks of the system under extreme scenarios; the threshold setting of rigid constraints is highly subjective and difficult to characterize marginal benefits. The introduction of risk measurement tools such as Conditional Value at Risk (CVaR) is mostly limited to the operation and scheduling level, failing to embed risk aversion awareness into the higher-level decision-making of capacity planning.

[0044] Deficiencies at the decision-making level. Most existing planning models presuppose equipment types and only optimize a single dimension such as the capacity of a single unit or the number of units installed. This results in limited alignment between the optimization results and actual engineering conditions, making it difficult to guide modular configuration decisions in engineering practice.

[0045] To this end, this application provides a method for configuring the capacity of a green electricity-to-hydrogen integrated energy system under uncertain supply and demand conditions. Figure 1 A flowchart illustrating a method for configuring the capacity of a green electricity-to-hydrogen integrated energy system under supply and demand uncertainty, according to one embodiment of this application, is shown. Figure 1 As shown, the technical solution provided in this application includes the following steps:

[0046] S1. Obtain meteorological data, load demand data, and technical and economic parameters of each candidate device for the target area.

[0047] According to one implementation method, meteorological data includes wind speed, solar irradiance, ambient temperature, etc.; load demand data includes time-series data of electrical load, heat load, and hydrogen load; and the technical and economic parameters of each candidate device include investment cost, maintenance cost, efficiency, service life, and upper and lower capacity limits.

[0048] S2. Based on meteorological data and load demand data, construct the physical architecture of the green electricity-to-hydrogen integrated energy system, and establish mathematical models of each device based on the physical architecture.

[0049] According to one implementation, the physical architecture of the green electricity-to-hydrogen integrated energy system consists of an energy supply side, a conversion side, an energy storage side, and a load side, and achieves bidirectional energy flow through interaction with the external power grid and the external hydrogen market.

[0050] The energy supply side includes wind power generation equipment, photovoltaic power generation equipment, and solar thermal equipment. Wind power generation equipment uses wind energy to drive the rotation of wind turbines to generate electricity; photovoltaic power generation equipment uses the photovoltaic effect of semiconductor materials to convert solar radiation energy into electrical energy; solar thermal equipment converts solar radiation energy into heat energy through a heat collection field, which is directly supplied to the heat load or stored in a thermal storage tank.

[0051] The conversion side includes an electrolyzer and a hydrogen gas turbine. When renewable energy is abundant, the electrolyzer converts electricity into hydrogen and recovers waste heat; when electricity is scarce, the hydrogen gas turbine consumes hydrogen for combined heat and power (CHP), simultaneously outputting electricity and high-quality waste heat. Together, they establish a two-way conversion channel between electricity, hydrogen, and heat.

[0052] Energy storage includes batteries, hydrogen storage tanks, and thermal storage tanks. Batteries are used to smooth out power fluctuations ranging from seconds to hours; hydrogen storage tanks are used for hydrogen storage across days or even seasons; and thermal storage tanks are used for thermal regulation ranging from hours to days. These three types of energy storage complement each other on a time scale.

[0053] The load side includes three types of end-user energy demand: electrical load equipment, thermal load equipment, and hydrogen load equipment. The system achieves bidirectional flow of electricity with the external public power grid through the grid connection point, and bidirectional trading of hydrogen with the regional hydrogen market through hydrogen pipelines or transport vehicles.

[0054] Furthermore, mathematical models of each device are established based on the physical architecture, including the operational constraint equations of each device, as well as the power balance equations, thermal power balance equations, and hydrogen mass balance equations of the green electricity hydrogen production integrated energy system established based on the operational constraint equations of each device.

[0055] Among them, the wind power generation equipment model: set the time period The real-time wind speed is Output of a single wind turbine Represented using a piecewise function:

[0056] ;

[0057] in, To cut into wind speed, Rated wind speed, To cut off the wind speed, This refers to the rated power of a single device.

[0058] If the system is installed Typhoon power generation equipment, during the period Total output of wind power for:

[0059] ;

[0060] in, The overall efficiency of the equipment's mechanical and electrical systems.

[0061] Photovoltaic power generation equipment model: setting time period The light intensity is The ambient temperature is Photovoltaic module panel temperature for:

[0062] ;

[0063] in, This refers to the battery's nominal operating temperature.

[0064] Output of a single photovoltaic module for:

[0065] ;

[0066] in, The area of ​​a single photovoltaic module. The photoelectric conversion efficiency is under standard test conditions. The temperature power coefficient, This is a reference temperature.

[0067] If the system is installed Block photovoltaic modules, then time period Total photovoltaic output for:

[0068] ;

[0069] in, This refers to the inverter efficiency.

[0070] Photothermal equipment model: setting time period The direct solar irradiance is The effective light-receiving area of ​​the solar collector is The power of the heat collection field is... for:

[0071] ;

[0072] in, The photothermal conversion efficiency of the solar collector.

[0073] A portion of the collected heat energy is directly supplied to the terminal heat load. The other part is stored in a thermal storage tank. ,satisfy:

[0074] .

[0075] Electrolyzer Model: Assume the system starts from... Choose from various electrolytic cell technologies. Introduce a binary variable. Indicate whether to select technology Electrolyzer configuration decisions include single-unit rated power. With the number of units installed ,technology Total installed capacity for:

[0076] .

[0077] Investment cost of electrolytic cells for:

[0078] ;

[0079] in, For the fixed investment costs related to the selection, This refers to the investment cost per unit capacity.

[0080] Hydrogen production rate of electrolyzer With heat production power They are respectively:

[0081] ;

[0082] ;

[0083] in, For input electrical power, For load rate, The hydrogen production efficiency varies with the loading rate. The heat production coefficient is denoted as .

[0084] Hydrogen gas turbine model: Total installed capacity of hydrogen gas turbine for:

[0085] ;

[0086] in, Model Number of installations This refers to the rated power of a single unit.

[0087] Its power generation capacity With hydrogen consumption rate The relationship is:

[0088] ;

[0089] in, Hydrogen consumption rate per unit of electricity generated.

[0090] Its heat generation power The relationship with power generation is as follows:

[0091] ;

[0092] in, This is the thermoelectric ratio.

[0093] Battery model: The rated energy capacity of the battery is Maximum charging and discharging power is The relationship between the two is as follows:

[0094] ;

[0095] in, This represents the maximum charge / discharge rate.

[0096] Its dynamic change in state of charge follows the energy balance equation:

[0097] ;

[0098] in, For time period The amount of stored energy, , These represent charge and discharge efficiencies, Self-discharge rate The time interval is the step size.

[0099] Hydrogen storage tank model: The dynamic change of hydrogen storage in the hydrogen storage tank follows the mass balance equation:

[0100] ;

[0101] in, For time period Hydrogen storage capacity, , These represent the hydrogen charging and dehydrogenation rates, respectively. , These are the charge / discharge efficiencies, Leakage rate.

[0102] Thermal storage tank model: Dynamic equations for the thermal energy storage level of the thermal storage tank:

[0103] ;

[0104] in, For time period Heat storage, , These are the rates of heat charging and heat release, respectively. , These are the charge / discharge efficiencies, This represents the heat loss rate.

[0105] Electric power balance equation:

[0106] ;

[0107] in, and These represent the power purchased and the power sold, respectively. and These are the battery discharge and charging power, respectively. For electrical load, For scene set, This is the set of scheduling time periods.

[0108] Heat power balance equation:

[0109] ;

[0110] in, , , The heat supplied directly to the load by the solar thermal equipment, electrolyzer, and hydrogen gas turbine, respectively. and These are the heat release and heat charging power of the thermal storage tank, respectively. This refers to the heat load.

[0111] Hydrogen mass balance equation:

[0112] ;

[0113] in, The hydrogen production rate of the electrolyzer. and These represent the rates of hydrogen purchase and sale, respectively. and These represent the hydrogen release and filling rates of the hydrogen storage tank, respectively. The hydrogen consumption rate of a hydrogen gas turbine. Hydrogen loading.

[0114] S3. Based on the historical statistical characteristics of meteorological data and load demand data, determine the uncertain parameter vector, generate a typical scenario set and its initial probability distribution based on the uncertain parameter vector, and construct a discrete probability fuzzy set based on norm constraints based on the typical scenario set.

[0115] According to one implementation method Figure 2 A diagram illustrating the framework for constructing a sub-Bruker fuzzy set according to an embodiment of this application is shown. See also: Figure 2 The cumulative distribution function of each random variable is fitted based on historical data of an uncertain parameter vector. Latin hypercube sampling is used to generate an initial scene set: the range of each random variable is divided into N intervals with equal probability. Representative points are randomly selected within each interval, and sampled values ​​are obtained through inverse transformation sampling. The N sampled values ​​are randomly arranged to generate N initial scenes, each with an initial probability of 1 / N. Through this method, N initial scenes are generated, forming the initial scene set. , This is the Nth initial scene.

[0116] The initial scene set is reduced using a synchronous back-substitution elimination method. In each iteration, the probabilistic distance loss between each scene and its nearest scene is calculated. The scene with the smallest probabilistic distance loss is deleted, and the probability of the deleted scene is added to that of its nearest scene. This process continues until the number of remaining scenes reaches a preset target value K, resulting in a typical scene set. and its initial probability distribution .

[0117] During the reduction process, scenarios where the initial concentrated renewable energy output is lower than a first preset threshold or where the load demand is higher than a second preset threshold are pre-identified as extreme scenarios, and these extreme scenarios are forcibly retained during the reduction process. The first and second preset thresholds can be set according to actual circumstances.

[0118] Furthermore, in the typical scenario set, among all probability distributions that satisfy the nonnegativity and normalization conditions, the 1-norm distance between them and the initial probability distribution does not exceed the first radius and - A set of probability distributions whose norm distance does not exceed the second radius. The first radius controls the total probability offset of all scenarios in the typical scenario set, and the second radius controls the maximum probability offset of a single scenario in the typical scenario set.

[0119] Specifically, consider any probability distribution on the typical scenario set S. ,satisfy and Discrete probabilistic fuzzy sets Defined as:

[0120] ;

[0121] in, It is a scene set Any probability distribution on the surface, where k is the index of the typical scenario set; Let the first radius be , This is the second radius.

[0122] S4. Based on the mathematical models of each device, the technical and economic parameters of each candidate device, the typical scenario set, and the discrete probability fuzzy set, construct a two-stage sub-Blu-ray capacity configuration optimization model.

[0123] According to one implementation, the two-stage distributed capacity configuration optimization model includes an upper-level capacity planning model and a lower-level operation scheduling optimization model.

[0124] The upper-level capacity planning model takes the mathematical models of each device and the technical and economic parameters of each candidate device as inputs, aims to minimize the sum of investment cost, maintenance cost and expected operating cost over the entire life cycle of the system, and uses equipment selection variables, installation quantity variables and single-unit capacity variables as decision variables to generate a capacity configuration scheme.

[0125] upper-level objective function for:

[0126] ;

[0127] in, The decision variables for the first stage include binary variables for equipment selection, integer variables for the number of units installed, and continuous variables for the capacity of a single unit. The feasible domain for equipment investment decisions; Investment cost (equivalent annual value); To maintain costs; Operating costs; Let $\frac{ ...

[0128] The upper-level model must meet constraints on the number of equipment installed, upper and lower limits of capacity, investment funds, policy compliance, spatial resources, and safety and energy efficiency.

[0129] Equipment Installation Quantity Constraints: Considering the park's spatial layout and policy restrictions, the installation quantity of each type of equipment must be within the allowable range. Let the number of installations for equipment i be... The minimum and maximum number of units allowed are respectively , ,but:

[0130] ;

[0131] in, This is the set of candidate device types.

[0132] Equipment capacity upper and lower limits: The rated capacity of a single piece of equipment or the total installed capacity must be within the technically permissible range. Let the rated capacity of equipment i be... The total installed capacity is If the technical specifications specify upper and lower limits for single-unit capacity and upper and lower limits for total installed capacity... , and the upper and lower limits of total installed capacity , ,but:

[0133] .

[0134] Investment capital constraints: Total investment shall not exceed the maximum investment capacity that the park can bear. .

[0135] ;

[0136] in, For fixed investment costs independent of capacity, This refers to the investment cost per unit capacity.

[0137] Policy compliance constraints include:

[0138] Renewable energy power consumption responsibility weight constraint: The proportion of annual renewable energy power consumption in the park to the total annual electricity consumption should not be less than the prescribed value. Assume the annual power generation from wind power and solar power is... , The park's total annual electricity consumption is ,but:

[0139] ;

[0140] Green electricity direct connection project access conditions: If the park adopts the green electricity direct connection model, the proportion of self-generated and self-consumed electricity from new energy sources shall not be less than 60% of the total available power generation and shall not be less than 30% of the total electricity consumption.

[0141] Constraints on the proportion of photovoltaic installation in new buildings: The area of ​​rooftop solar photovoltaic installation in new industrial buildings shall not be less than the preset proportion of the actual installation area, usually taken as 50% to 70%.

[0142] Spatial resource constraints include:

[0143] Wind farm site area constraints: Assume the rated power of a single wind turbine is... The empirical coefficient for the area occupied per unit capacity is: The maximum area available for wind farm development in the park is ,but:

[0144] ;

[0145] in, This is a set of wind turbine types.

[0146] Photovoltaic installation capacity limit constraint: Let the effective total area of ​​the park where photovoltaics can be installed be... The installed capacity per unit area is Total installed photovoltaic capacity satisfy:

[0147] ;

[0148] Land constraints for hydrogen production and hydrogen gas turbine equipment: Let the single-unit floor area coefficient of equipment i be... The total land area available for the installation of this type of equipment in the park is ,but:

[0149] ;

[0150] in, This is the set of equipment types that have floor space constraints.

[0151] Safety distance constraints for hydrogen energy facilities: To ensure safety, hydrogen production equipment and storage tanks must meet minimum fire safety distance requirements between themselves and surrounding buildings and structures. This constraint typically needs to be discretized in conjunction with the specific site layout.

[0152] Safety and energy efficiency constraints:

[0153] System annual comprehensive energy efficiency The ratio of annual useful energy output to annual primary energy input must not be lower than the minimum standard. :

[0154] ;

[0155] in, For day d, For the t-th scheduling period, This represents the total number of daily scheduling periods. The actual power generation of the wind turbine during time period t. The actual power generation of the photovoltaic power generation equipment during time period t. For the scheduling time interval step, , , These are the actual load values ​​for electricity, heat, and hydrogen, respectively. Hydrogen has a low calorific value. This refers to externally purchased electrical power.

[0156] These constraints together constitute the complete feasible region of the upper-level capacity planning, ensuring the rigor and engineering applicability of the model.

[0157] The lower-level operation scheduling optimization model takes the capacity configuration scheme, typical scenario set and mathematical model of each device as input, takes minimizing the energy shortage under each typical scenario as the objective, and takes the hourly operating power of each device as the decision variable to solve for the minimum energy shortage and minimum operating cost under each typical scenario. The minimum operating cost is then fed back to the upper-level capacity planning model to calculate the expected operating cost.

[0158] The energy shortage is the sum of the equivalent energy calculated from the power shortage of electrical load, the power shortage of thermal load, and the rate of power shortage of hydrogen load within the dispatch cycle.

[0159] For each typical scenario The lower-level objective function is:

[0160] ;

[0161] in, Time periods The shortage of electricity, heat, and hydrogen load; The lower heating value of hydrogen is used to convert hydrogen shortage into equivalent energy shortage. These are the weighting coefficients; For running schedule variables; This is the feasible domain for operation.

[0162] The lower-level model also outputs the minimum daily operating cost for this scenario. :

[0163] ;

[0164] in, , These are the electricity purchase and sales prices for time period t, respectively, following the time-of-use pricing mechanism; , These refer to the purchase price of hydrogen and the selling price of hydrogen. , These represent the electricity purchased and sold during time period t, respectively. , These represent the hydrogen purchase and sales rates for time period t, respectively.

[0165] The lower-level model must satisfy the following constraints:

[0166] The equipment operation constraints and energy balance constraints are specifically the same as the operation constraint equations for each piece of equipment in S2, as well as the electrical power balance equation, thermal power balance equation, and hydrogen matter balance equation.

[0167] Constraints on interaction with the external network include limits on electricity purchase and sale and limits on hydrogen purchase and sale rates.

[0168] , ;

[0169] , ;

[0170] in, , These represent the electricity purchased and sold during time period t, respectively. , These are the maximum power purchase and power sales limits, respectively. , These represent the hydrogen purchase and sales rates for time period t, respectively. , These are the maximum hydrogen purchase and sales rate limits, respectively.

[0171] Non-negative constraint on supply shortage:

[0172] ;

[0173] in, , , These represent the power shortage of electricity and heat load and the rate of hydrogen load shortage during time period t, respectively.

[0174] Energy storage cycle start and end constraints:

[0175] , , ;

[0176] in, , These represent the state of charge of the battery at the beginning and end of the scheduling cycle. , These represent the hydrogen storage capacity of the hydrogen storage tank at the beginning and end of the scheduling cycle, respectively. , These represent the amount of heat stored in the thermal storage tank at the beginning and end of the scheduling cycle. To ensure the long-term cyclical operation of the energy storage equipment, the initial and final states of each energy storage unit must remain consistent.

[0177] Building thermodynamic constraints: The dynamic relationship between system heating power and building indoor temperature is described by a first-order ETP model:

[0178] ;

[0179] in, The indoor temperature for the next time period t+1. Let be the indoor temperature during time period t. The heating power supplied by the system to the building during time period t. Equivalent heat capacity of a building reflects the building envelope's ability to absorb and store heat. The equivalent thermal resistance of a building reflects the thermal insulation performance of its building envelope. The outdoor ambient temperature during time period t;

[0180] Indoor temperature should be maintained within the range of human thermal comfort.

[0181] ;

[0182] in, The lower limit of indoor temperature, This is the upper limit of indoor temperature;

[0183] To ensure the sustainable adjustment of thermal inertia during the scheduling cycle, the initial and final values ​​of the indoor temperature must be consistent:

[0184] ;

[0185] in, The indoor temperature at the start of the scheduling cycle.

[0186] The lower-level model provides key inputs to the upper-level model: for a given capacity decision x, the lower-level model solves for the optimal energy shortage and the minimum daily operating cost under the given scenario. Then, the energy shortage ENS value is used to construct the Conditional Value at Risk (CVaR) constraint, and the annualized operating cost constitutes the expected operating cost term in the upper-level objective function.

[0187] The upper-level decision x affects the upper and lower limits of equipment capacity and energy storage capacity, thereby changing the feasible domain of lower-level operation and thus affecting the values ​​of ENS and operating costs. This coupling relationship is reflected through a two-level programming model, realizing coordinated optimization from both economic and physical perspectives.

[0188] S5. Based on the discrete probability fuzzy set, conditional risk value constraints are introduced into the two-stage sub-Bruker capacity allocation optimization model to obtain a two-stage sub-Bruker capacity allocation optimization model with conditional risk value constraints.

[0189] According to one implementation, to ensure the system's reliability under the worst-case probability distribution, a conditional risk value constraint based on a split-bar framework is introduced. For a given capacity decision x, a random variable is defined. For the system under uncertain parameters The minimum energy shortage. Its conditional value at risk is:

[0190] ;

[0191] in, For a probability distribution P, where there is a lack of energy supply The conditional value of risk at confidence level α, where α∈(0,1) is the confidence level, reflects the degree to which decision-makers averse to extreme risks. The threshold for value at risk. Let ENS be the expected value under probability distribution P. The requirement is that, for all possible distributions within the fuzzy set D, the CVaR of ENS does not exceed a given threshold. ,Right now:

[0192] ;

[0193] This constraint ensures that even under the worst probability distribution, the tail risk of power shortage in the system is kept within an acceptable range.

[0194] To transform the formulas into a computable constraint form, a discrete scene set is used. and its initial probability Distribution. For any probability distribution p∈D, CVaR can be expressed as:

[0195] ;

[0196] Where R represents the set of real numbers.

[0197] in, , , To determine the optimal energy supply for the given scenario, the constraint is equivalent to:

[0198] ;

[0199] By employing dual transformation or directly introducing auxiliary variables, this constraint can be transformed into a linear constraint form. This constraint is then directly added to the upper-level capacity planning model as an additional constraint that the decision variable x must satisfy, thus obtaining the two-stage sub-Bruker capacity allocation optimization model with conditional value-at-risk constraints.

[0200] S6. Use the column and constraint generation algorithm to solve the two-stage sub-Bruker capacity allocation optimization model with conditional risk value constraints. Iterate until the convergence condition is met and output the optimal capacity allocation scheme.

[0201] According to one implementation method Figure 3 A flowchart illustrating the column and constraint generation algorithm solution according to an embodiment of this application is shown. See also: Figure 3 S6 includes the following sub-steps:

[0202] S61. Set the worst-case scenario set to an empty set, set the upper bound to positive infinity, and set the lower bound to negative infinity.

[0203] S62. Under the current worst-case scenario set, solve the upper-level capacity planning model, and use the deterministic equivalent form of the conditional risk value constraint under the current worst-case scenario set as an additional constraint to obtain the current capacity configuration scheme and update the lower bound.

[0204] To mathematically represent the above iterative process, the column and constraint generation (C&CG) algorithm is used to reconstruct the original problem into a main problem and subproblems.

[0205] The main problem is to optimize capacity decisions under the current worst-case scenario set, and introduces a deterministic equivalent form of CVaR constraints. Let L be the index set of the current worst-case scenarios, then the main problem is:

[0206] ;

[0207] satisfy:

[0208] ;

[0209] ;

[0210] ;

[0211] ;

[0212] ;

[0213] ;

[0214] in, For the scene The initial probability weights are used to calculate the expected running cost; For the scene The initial probability is used for CVaR constraints. For the first The set of lower-level runtime scheduling decision variables for each scenario. For the first Energy shortage in certain scenarios For the first Daily operating costs in each scenario Decision for a given capacity and uncertain parameters The next lower-level feasible region. Solving the main problem yields... and the lower bound LB.

[0215] S63. Under the current capacity configuration scheme, solve the lower-level operation scheduling optimization model for each typical scenario to obtain the minimum energy shortage for each typical scenario; based on the minimum energy shortage for each typical scenario, search for the probability distribution that maximizes the conditional risk value in the discrete probability fuzzy set to obtain the worst-case probability distribution and the maximum conditional risk value, and update the upper bound according to the expected operating cost under the current capacity configuration scheme and the worst-case probability distribution.

[0216] Subproblems are for a given Search for the probability distribution that maximizes CVaR in the fuzzy set D:

[0217] ;

[0218] First, we need to solve the underlying execution problem for each scenario k to obtain... Subsequently, Substitute the values ​​and solve the following linear programming problem:

[0219] ;

[0220] satisfy:

[0221] ;

[0222] ;

[0223] ;

[0224] ;

[0225] in, The optimal capacity configuration scheme obtained by solving the main problem in the current iteration step is used to solve the problem and obtain the maximum conditional value of risk. and worst-case probability distribution .

[0226] S64. If the difference between the upper and lower bounds is less than the convergence accuracy, then the current capacity configuration scheme is the optimal solution. Stop the iteration and output the optimal capacity configuration scheme. If the difference between the upper and lower bounds is greater than or equal to the convergence accuracy, then select the scenario with the largest energy shortage from the typical scenario set and add it to the worst scenario set. Return to S62 and execute again.

[0227] like ,in, To achieve convergence accuracy, the current If the reliability requirements are met, the algorithm terminates and outputs the optimal capacity configuration. Otherwise, it selects the scenario that maximizes the capacity. Add the worst-case scenario set L and return to the S62 master problem to solve it again, iterating until convergence.

[0228] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of this application may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification. It should be understood that, in order to simplify this disclosure and aid in understanding one or more aspects of the various applications, various features of this application are sometimes grouped together in a single embodiment, figure, or description thereof in the above description of exemplary embodiments of this application. However, this approach should not be construed as reflecting an intention that the claimed application requires more features than are expressly recited in each claim. Furthermore, unless otherwise specified, the use of ordinal numbers “first,” “second,” “third,” etc., to describe ordinary objects merely indicates different instances of similar objects and is not intended to imply that such described objects must have a given temporal, spatial, or sequential order or in any other manner.

[0229] Although this application has been described with reference to a limited number of embodiments, those skilled in the art will understand from the foregoing description that other embodiments are conceivable within the scope of this application described herein. Furthermore, it should be noted that the language used in this specification has been chosen primarily for readability and edibility purposes, and not for interpreting or limiting the subject matter of this application. Therefore, many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the appended claims. The disclosure of this application is illustrative and not restrictive, and the scope of this application is defined by the appended claims.

Claims

1. Capacity configuration methods for integrated green electricity-to-hydrogen energy systems under supply and demand uncertainties, including: S1. Obtain meteorological data, load demand data, and technical and economic parameters of each candidate device for the target area; S2. Based on the meteorological data and the load demand data, construct the physical architecture of the green electricity-to-hydrogen integrated energy system, and establish mathematical models of each device based on the physical architecture. S3. Based on the historical statistical characteristics of the meteorological data and the load demand data, determine the uncertain parameter vector, generate a typical scenario set and its initial probability distribution based on the uncertain parameter vector, and construct a discrete probability fuzzy set based on norm constraints based on the typical scenario set. S4. Based on the mathematical models of each device, the technical and economic parameters of each candidate device, the typical scenario set, and the discrete probability fuzzy set, construct a two-stage sub-Blu-ray capacity configuration optimization model. S5. Based on the discrete probability fuzzy set, conditional risk value constraints are introduced into the two-stage sub-Brussels capacity configuration optimization model to obtain a two-stage sub-Brussels capacity configuration optimization model with conditional risk value constraints. S6. Use the column and constraint generation algorithm to solve the two-stage sub-Bruker capacity allocation optimization model with conditional risk value constraints, iterate until the convergence condition is met, and output the optimal capacity allocation scheme.

2. The method as described in claim 1, wherein, In S2, the equipment in the physical architecture includes power supply equipment, conversion equipment, energy storage equipment, and load equipment; the power supply equipment includes wind power generation equipment, photovoltaic power generation equipment, and solar thermal equipment; the conversion equipment includes electrolyzers and hydrogen gas turbines; the energy storage equipment includes batteries, hydrogen storage tanks, and thermal storage tanks; the load equipment includes electrical load equipment, thermal load equipment, and hydrogen load equipment; the mathematical model of each device includes the operating constraint equations of each device, as well as the electrical power balance equation, thermal power balance equation, and hydrogen mass balance equation of the green electricity to hydrogen integrated energy system established based on the operating constraint equations of each device.

3. The method as described in claim 1, wherein, In step S3, generating a typical scenario set and its initial probability distribution based on the uncertain parameter vector specifically includes: The cumulative distribution function of each random variable is fitted based on the historical data of the uncertain parameter vector; Latin hypercube sampling is used to generate an initial scene set: the range of values ​​of each random variable is divided into N intervals with equal probability, representative points are randomly selected in each interval, and sampled values ​​are obtained through inverse transformation sampling. The N sampled values ​​are randomly arranged to generate N initial scenes, and the initial probability of each initial scene is 1 / N. The initial scene set is reduced using a synchronous back-substitution elimination method. In each iteration, the probability distance loss between each scene and its nearest scene is calculated, and the scene with the smallest probability distance loss is deleted. The probability of the deleted scene is added to its nearest scene until the number of remaining scenes reaches a preset target value, thus obtaining the typical scene set and its initial probability distribution.

4. The method of claim 3, wherein, During the process of reducing the initial scenario set using the synchronous back-substitution elimination method, scenarios in the initial scenario set where the renewable energy output is lower than a first preset threshold or the load demand is higher than a second preset threshold are identified in advance as extreme scenarios, and these extreme scenarios are forcibly retained during the reduction process.

5. The method of claim 1, wherein, In S3, the discrete probability fuzzy set based on norm constraints is: In the typical scenario set, among all probability distributions that satisfy the non-negativity and normalization conditions, the set of probability distributions whose 1-norm distance from the initial probability distribution does not exceed the first radius and whose ∞-norm distance does not exceed the second radius; the first radius is used to control the total probability offset of all scenarios in the typical scenario set, and the second radius is used to control the maximum probability offset of a single scenario in the typical scenario set.

6. The method of claim 1, wherein, In S4, the two-stage distributed capacity configuration optimization model includes an upper-level capacity planning model and a lower-level operation scheduling optimization model. The upper-level capacity planning model takes the mathematical models of each device and the technical and economic parameters of each candidate device as inputs, aims to minimize the sum of investment cost, maintenance cost and expected operating cost over the entire life cycle of the system, and uses equipment selection variables, installation quantity variables and single-unit capacity variables as decision variables to generate a capacity configuration scheme. The lower-level operation scheduling optimization model takes the capacity configuration scheme, the typical scenario set and the mathematical model of each device as input, takes minimizing the energy shortage under each typical scenario as the objective, and takes the hourly operating power of each device as the decision variable to solve for the minimum energy shortage and minimum operating cost under each typical scenario. The minimum operating cost is then fed back to the upper-level capacity planning model to calculate the expected operating cost. The energy shortage is the sum of the equivalent energy calculated from the power shortage of electrical load, the power shortage of thermal load, and the rate shortage of hydrogen load within the scheduling cycle.

7. The method of claim 6, wherein, In S5, the conditional value at risk constraint is expressed as follows: Under all possible probability distributions within the discrete probability fuzzy set, the conditional risk value of insufficient energy output by the lower-level operation scheduling optimization model does not exceed a preset threshold. The conditional value at risk is used to measure the expected level of the energy shortage at the tail of the probability distribution.

8. The method as described in claim 7, wherein in step S6, the column and constraint generation algorithm is used to solve the two-stage sub-Bruker capacity allocation optimization model with conditional risk value constraints, specifically including: S61. Set the worst-case scenario set to an empty set, set the upper bound to positive infinity, and set the lower bound to negative infinity; S62. Under the current worst-case scenario set, solve the upper-level capacity planning model, and use the deterministic equivalent form of the conditional risk value constraint under the current worst-case scenario set as an additional constraint to obtain the current capacity configuration scheme and update the lower bound. S63. Under the current capacity configuration scheme, solve the lower-level operation scheduling optimization model for each typical scenario to obtain the minimum energy shortage for each typical scenario; Based on the minimum energy shortage in each typical scenario, the probability distribution that maximizes the conditional risk value is searched within the discrete probability fuzzy set to obtain the worst-case probability distribution and the maximum conditional risk value. The upper bound is then updated based on the current capacity configuration scheme and the expected operating cost under the worst-case probability distribution. S64. If the difference between the upper bound and the lower bound is less than the convergence accuracy, then the current capacity configuration scheme is the optimal solution, stop the iteration and output the optimal capacity configuration scheme; if the difference between the upper bound and the lower bound is greater than or equal to the convergence accuracy, then select the scenario with the largest energy shortage from the typical scenario set and add it to the worst scenario set, and return to S62 to re-execute.

9. A computing device, comprising: At least one processor; as well as A memory storing program instructions, wherein the program instructions are configured to be executed by the at least one processor, the program instructions including instructions for performing the method as described in any one of claims 1-8.

10. A readable storage medium storing program instructions that, when read and executed by a computing device, cause the computing device to perform the method as described in any one of claims 1-8.

Images