Hydrogen-containing mixed energy supply infrastructure planning method

Abstract

The application discloses a hydrogen-containing mixed energy supply infrastructure planning method, and relates to the field of hydrogen energy, and comprises the following steps: step 1, three-network coupling modeling of an electric network, a hydrogen network and a traffic network: coupling network physical scene construction and optimization target establishment, traffic network constraint construction, hydrogen network constraint construction, electric network constraint construction and coupling constraint construction; step 2, decision dependence distribution robust optimization model construction: hydrogen demand decision dependence fuzzy set construction and distribution robust optimization framework construction; step 3, distribution robust optimization problem reconstruction and solving method: decision dependence scene probability construction, norm item linearization and opportunity constraint reconstruction; and step 4, three-network coupling system example simulation analysis. The application effectively solves the problem of whether a hydrogen vehicle or a hydrogen refueling station is first developed when hydrogen fuel vehicles have not been popularized, realizes reasonable planning of hydrogen infrastructure, reduces investment and operation costs, improves new energy consumption of an electric network and effectively avoids energy waste in the early development of hydrogen vehicles.

Description

Technical Field

[0001] This invention relates to the field of hydrogen energy, and more particularly to a method for planning hydrogen-containing hybrid energy supply infrastructure. Background Technology

[0002] In recent years, hydrogen, as a clean, zero-carbon fuel, has received increasing attention. Hydrogen fuel cell vehicles are considered a new type of transportation to replace traditional fuel vehicles. In terms of driving range and refueling time, hydrogen fuel cell vehicles are significantly superior to electric vehicles, and they possess characteristics such as high energy density, zero emissions, and rapid refueling, showing broad application prospects. However, due to the immaturity of hydrogen fuel cell vehicle technology and the current low proportion of hydrogen refueling stations, hydrogen fuel cell vehicles have not yet achieved widespread adoption. How to accelerate the popularization of hydrogen fuel cell vehicles in the absence of hydrogen infrastructure, and how to attract investment in hydrogen refueling stations when there are no hydrogen fuel cell vehicles on the road, is known as the "chicken and egg" problem. However, before the hydrogen fuel cell vehicle market can sell cars to the public, at least some hydrogen refueling stations must be operational. The profitability of hydrogen refueling stations largely depends on utilization rates, and low utilization rates are inevitable in the early stages of market development. How to best coordinate the development of infrastructure and vehicle sales is crucial for a successful transition from traditional energy vehicles to hydrogen fuel cell vehicles. Therefore, before hydrogen fuel cell vehicles become widespread, hydrogen refueling infrastructure needs to be planned and constructed. However, without precise data on future hydrogen refueling demand for vehicles, how can hydrogen refueling stations be rationally planned while ensuring socio-economic benefits? Furthermore, regarding the synergistic effects of hydrogen networks, power grids, and transportation networks, can the introduction of hydrogen infrastructure improve the utilization rate of renewable energy and alleviate traffic congestion? No effective solutions to these problems have yet been proposed.

[0003] Previous inventions have extensively studied the joint planning problem of power and hydrogen infrastructure considering source-load uncertainty. Most of these methods treat hydrogen demand as an exogenous uncertainty, solving it with robust optimization methods, or using stochastic optimization methods when the exact probability distribution is known. However, none of these methods account for the limitations of hydrogen refueling station planning at the current stage, when hydrogen fuel cell vehicles are not yet widespread. The resulting planning may lead to low energy utilization, over-planning, or under-planning. In fact, the larger the investment in hydrogen infrastructure, the more consumers will be attracted to purchase hydrogen fuel cell vehicles; that is, the development of the hydrogen fuel cell vehicle market depends on the planning decisions for hydrogen refueling stations. Conversely, the uncertainty of hydrogen refueling demand for hydrogen fuel cell vehicles also needs to be considered in the hydrogen refueling station planning problem. Therefore, this constitutes a decision dependency relationship between hydrogen demand and hydrogen refueling station planning decisions.

[0004] Based on the above discussion, this invention argues that in the early stages of hydrogen fuel cell vehicle development, the decision-dependent uncertainty of hydrogen refueling demand needs to be considered in hydrogen infrastructure planning. In reality, many stochastic factors are influenced by decision choices, a phenomenon known as decision-dependent uncertainty (DDU). For example, production decisions are influenced by investment information, component reliability by maintenance decisions, and transportation demand by road expansion decisions. Depending on the specific decision-making mechanism in the application scenario, decision-dependent uncertainty can be broadly categorized into two types: 1) The realization of uncertain factors affecting decision-making, which often models the optimization problem as robust or stochastic optimization. 2) The probability distribution of uncertain factors affecting decision-making, whose realization is independent of the decision. The former has been thoroughly studied in power grid scenarios, often utilizing optimality conditions and projection theory in solving optimization problems. The latter is typically modeled as a distributed robust optimization (DRO) problem, applied to pre-disaster planning, system or road maintenance decision-making, etc. However, based on previous research experience, robust optimization models tend to make relatively conservative optimization decisions, while stochastic optimization requires obtaining accurate probability distributions of uncertain parameters in advance. In practical problems, we may only be able to obtain partial data on the hydrogen refueling requirements of hydrogen fuel cell vehicles, which happens to match the modeling characteristics of the partial Bruker optimization problem.

[0005] In decision-dependent blob optimization problems with uncertainty, the key to modeling is defining fuzzy sets. Widely used fuzzy sets include those based on moments and those based on Wasserstein metric. For moment-based decision-dependent blob optimization problems, the mean and variance of the uncertain parameters depend on the decision, but such fuzzy sets cannot guarantee any convergence from the unknown distribution to the true distribution. For Wasserstein-based decision-dependent blob optimization problems, constructing an empirical distribution using samples, without assigning probability weights, and quantifying the confidence upper limit by the Wasserstein radius, can achieve better out-of-sample performance. Furthermore, Wasserstein-based fuzzy sets can control the conservatism of the solution through the radius, thus ensuring the flexibility of power system operation. Moreover, Wasserstein-based fuzzy sets have a potential polyhedral structure, making them more suitable for algorithm development within linear programming or linear conical duality frameworks. Existing literature has limited research on decision-dependent uncertainty based on Wasserstein metric, only considering cases where both the decision and uncertain variables are binary variables, which does not align with the characteristics of hydrogen infrastructure planning problems. Furthermore, the theory of decision-dependent blob optimization based on Wasserstein-based fuzzy sets has not yet been applied in energy grids.

[0006] In summary, to accelerate the low-carbon development of urban transportation, and given the current lack of widespread adoption of hydrogen fuel cell vehicles, how to rationally plan hybrid energy supply infrastructure capable of simultaneous charging and refueling is a research gap in the field of hydrogen infrastructure planning. Simultaneously, it is necessary to fully utilize the synergistic effects of the three-network coupling system, through site selection and scale planning of infrastructure, to fully utilize excess renewable energy from the power grid and redistribute traffic flow in the transportation network. Furthermore, regarding modeling methods for robust optimization problems considering uncertainty, as discussed above, decision-dependent partial robust optimization with low decision conservatism has significant advantages in hydrogen infrastructure planning problems and has been almost entirely unexplored in energy grid scenarios.

[0007] Therefore, those skilled in the art are dedicated to developing a planning methodology for hydrogen-containing hybrid energy supply infrastructure (HESI), including the site selection and scale planning of fast charging stations (FCS), hydrogen refueling stations (HRS), power-to-gas (P2G) units, and hydrogen pipelines. While hydrogen fuel cell vehicles (HFCVs) are not yet widespread, this method can effectively reduce the conservatism of planning decisions, while also adapting well to the increasing demand for hydrogen refueling in the future. Furthermore, this method can effectively reduce photovoltaic (PV) curtailment rates and lower traffic congestion costs. This is of great significance for promoting the widespread adoption of hydrogen fuel cell vehicles. Summary of the Invention

[0008] In view of the above-mentioned deficiencies of the prior art, the technical problem to be solved by the present invention is that in the early stage of hydrogen fuel cell vehicle development, hydrogen vehicle development and hydrogen infrastructure planning are mutually constrained, the utilization rate of hydrogen infrastructure is low, and insufficient data on hydrogen refueling demand leads to planning difficulties, namely the "chicken or egg" problem between vehicles and stations. At the same time, most existing hydrogen-containing hybrid energy supply infrastructure planning technologies only consider energy-side benefits and do not fully utilize the synergistic effect of the coupled network of the power grid, hydrogen network, and transportation network to achieve joint optimization of the three networks.

[0009] To achieve the above objectives, the present invention provides a method for planning hydrogen-containing hybrid energy supply infrastructure, the method comprising the following steps:

[0010] Step 1: Model the coupling of the power grid, hydrogen grid, and transportation network;

[0011] Step 2: Constructing the decision-dependent sub-bar optimization model;

[0012] Step 3: Reconstruction and solution method of the Bruker optimization problem;

[0013] Step 4: Simulation analysis of a three-network coupled system;

[0014] Step 1 further includes:

[0015] Step 1.1: Constructing the physical scenario of the coupled network and establishing optimization objectives;

[0016] Step 1.2: Construction of transportation network constraints;

[0017] Step 1.3: Construction of hydrogen network constraints;

[0018] Step 1.4: Constructing power grid constraints;

[0019] Step 1.5: Constructing coupling constraints;

[0020] Step 2 also includes:

[0021] Step 2.1: Hydrogen demand decision depends on fuzzy set construction;

[0022] Step 2.2: Construct the Brussels Bar Optimization Framework;

[0023] Step 3 also includes:

[0024] Step 3.1: Constructing the probability of decision-dependent scenarios;

[0025] Step 3.2: Linearization of norm terms;

[0026] Step 3.3, Opportunity Constraint Restructuring.

[0027] Furthermore, step 1.1 also includes: in the planning of hydrogen-containing hybrid energy supply infrastructure, planning for hydrogen refueling stations, charging stations, electricity-to-gas devices, and hydrogen pipelines is carried out based on the charging and hydrogen refueling needs and economic benefits of the nodes, with investment decisions made as follows: and w l All variables are binary decision variables; assuming a fixed amount of hydrogen for each hydrogen fuel cell vehicle, the hydrogen load is a discrete variable; the hydrogen grid is coupled to the power grid via a power-to-gas (P2G) device, making full use of the grid's renewable energy. After hydrogen is produced by the water electrolysis device, it is compressed by a compressor and stored in a hydrogen storage tank (HS). After energy conversion and storage, it is supplied to hydrogen fuel cell vehicles (HFCVs) through hydrogen refueling stations (HRS); when PV output is in surplus, the portion that the energy storage device (ES) cannot store is produced by P2G to reduce the curtailment rate; when PV output is insufficient, the hydrogen load that P2G cannot meet is addressed.

[0028] The optimization objective is to minimize the sum of HESI's annual investment cost and annual average operating cost, as follows:

[0029]

[0030]

[0031]

[0032] Among them, c hy c P2G cHS , and c FCS The annual investment costs for HRS, P2G, HS, hydrogen pipelines, and FCS are respectively. and These represent the investment capacities of HRS, P2G, HS, and FCS at node i, respectively. For a set of nodes;

[0033] Operating costs include penalties for curtailment of solar power, electricity trading with the main grid, penalties for not meeting charging load requirements, traffic congestion, and penalties for not meeting hydrogen charging load requirements.

[0034] Furthermore, step 1.2 also includes: assuming a finite-range transportation network containing multiple starting points o and multiple ending points d, and multiple paths p that allow electric vehicles to travel from starting point o to ending point d, each path p consisting of multiple links l; establishing the link traffic at each moment in the transportation network. With path traffic Relational model:

[0035]

[0036]

[0037]

[0038] To ensure that each electric vehicle performs a charging operation once when passing through the transportation network, the following constraint is added, representing the flow on the link at the fast charging station connected to grid node i. With path traffic Relational model:

[0039]

[0040]

[0041]

[0042] According to the US Highway Bureau function, the relationship between the travel time of each vehicle on link l and the link traffic is as follows:

[0043]

[0044] The total congestion time on link l is:

[0045]

[0046] Referring to the traffic flow optimization management problem, according to the Wardrop user equilibrium principle, when the optimal decision is achieved in this optimization problem, no vehicle in the traffic network can reduce its total cost by changing its own driving decision; this optimization problem is equivalent to the following Karush-Kuhn-Tucker (KKT) conditions:

[0047]

[0048] Equations (12) and (11) are linearized using the Big M method and piecewise linearization method, respectively, and then replaced with equations (13) and (14), respectively:

[0049]

[0050]

[0051] Furthermore, step 1.3 also includes: for the HS constraint, equation (15) is the HS energy balance equation, equation (16) limits the maximum hydrogen injection and release of HS, and the state of charge and recycling conditions of HS are constrained by equations (17) and (18):

[0052]

[0053]

[0054]

[0055] H i,0 =H i,T (18)

[0056] The following limitations exist in planning decisions regarding HRS, P2G, and hydrogen pipelines:

[0057]

[0058]

[0059] To make the final optimization problem easier to handle, a hydrogen pipeline modeling method based on the concept of regions is applied; assuming that each candidate hydrogen node is analogous to a region and that the net flow of each region is balanced, the balance constraint (21) and the hydrogen flow rate limit (22) are obtained:

[0060]

[0061]

[0062] Meanwhile, the hydrogen pipelines have the capacity for coil storage, and the storage dynamic equations for each hydrogen pipeline are (23), (24) and (25), which respectively give the pipeline capacity constraints and cyclic conditions:

[0063]

[0064]

[0065]

[0066] The supply and demand balance constraint of the hydrogen network is represented by equation (26), which characterizes the conservation relationship of hydrogen production, hydrogen demand, pipeline transportation, storage, and purchase:

[0067]

[0068] Equation (27) constrains the upper and lower limits of hydrogen purchases at each node:

[0069]

[0070] In constraint (28), The variable representing the existence of a hydrogen pipeline connecting the hydrogen source and node i is a binary variable and requires additional constraints in specific simulation cases depending on the specific network construction.

[0071]

[0072] Constraint (29) ensures the nonnegativity of hydrogen abandonment demand:

[0073]

[0074] Constraint (30) specifies the upper and lower limits of the total hydrogen demand for the HFCV:

[0075]

[0076] In equation (30) It is an uncertain variable.

[0077] Furthermore, step 1.4 also includes: equations (31) and (32) for active power balance and line capacity constraints:

[0078]

[0079]

[0080] Constraint (33) limits the upper and lower limits of the amount of electricity each node can purchase from the main grid:

[0081]

[0082] Equations (34) and (35) represent the upper and lower limits of P2G capacity and PV waste, respectively:

[0083]

[0084]

[0085] Constraint (36) ensures the nonnegativity of the charging demand for abandoned EVs:

[0086]

[0087] Constraint (37) limits the total charging demand of EVs:

[0088]

[0089] Constraints (38)-(42) represent the relevant constraints of ES:

[0090]

[0091]

[0092]

[0093] E e,0 =E e,T (41)

[0094]

[0095] The actual output of the photovoltaic (PV) power is set as an exogenous uncertain variable, described by the uncertainty set V; where the predicted value of the PV output is... The fluctuation range is The threshold value for the uncertain variable is used to adjust the robustness of photovoltaic power output.

[0096]

[0097] Step 1.5 further includes: coupling constraints (44) and (45) representing the relationship between vehicle flow and charging demand and hydrogen refueling demand, respectively:

[0098]

[0099]

[0100] The hydrogen production of P2G is expressed as a steady-state linear function of the input power, as shown in equation (46):

[0101]

[0102] Furthermore, step 2.1 also includes: setting the uncertain variable hydrogen demand as... N bLet N be the total number of nodes, and N be the possible values ​​of the uncertain quantity at each node. s There are 100 scenes. The index is n; in DRO-DDU, it is generally assumed that the DDU set has finite decision-independent supports, because assigning a probability of 0 to a particular element is equivalent to excluding it from the supports, achieving a point-to-set mapping, where decision dependence is reflected by a varying probability mass function; for finite supports The candidate probability distribution in is p n (w hy ), corresponding scenarios And ||p n (w hy )||1=1;

[0103] Define Wasserstein metric as follows:

[0104]

[0105] in, To support the probability distribution set above Ξ, ∏ is the joint probability distribution of u1 and u2, with marginal distributions P1 and P2 respectively; ||·|| is an arbitrary norm, and ||u1-u2|| is the cost of moving a unit mass from distribution P1 to P2;

[0106] Fuzzy sets are defined as follows:

[0107]

[0108] This fuzzy set can be viewed as an empirical distribution. A sphere centered at r with radius r, where radius r can explicitly control the conservatism of the decision outcome;

[0109] Consider u t Experience distribution and decision-making hy Related, that is Among them, control u t Probability measure of ∈Ξ It is a decision w hy The function, and the sample space Ξ and w hy Irrelevant; the DDU fuzzy set is written in the following form:

[0110]

[0111] Since the hydrogen demand is discretely supported, the equivalent form of the above fuzzy set is:

[0112]

[0113] Furthermore, step 2.2 also includes: the investment optimization problem constructed in step one is equivalent to the following decomposed bar optimization.

[0114]

[0115] Where f(·) and h(·) are the investment and operating costs of HESI, respectively; combining the fuzzy set (50) proposed in step 2.1, the reconstructed form of the DDU-DRO model is derived:

[0116]

[0117]

[0118] ε≥0 (52c)

[0119]

[0120] Furthermore, step 3.1 also includes: as mentioned in step 2.1, the dimension of u is N. b The possible implementations of u are N s A discrete set of values, i.e., {0,2,...,N} s -1}, with index k, increasing from low to high; the number of scenes is The index is n; For the investment decision at node i; given the parameter vector and They are respectively The probabilities of the possible values ​​of hydrogen charging demand at node i when taking values ​​of 0 and 1 are obtained from historical data, and there is an internal relationship between them. and The probability of scenario n at node i is:

[0121]

[0122] in, Then the decision depends on the probability distribution The probability of scenario n is:

[0123]

[0124] The probability expression (54) is a high-dimensional nonlinear function of the decision variable, and the number of scenarios is exponentially related to the number of nodes. Using an improved distribution shaping theory, a set of linear constraints is used to characterize the scenario probabilities related to the decision. By introducing a truncation vector, a linear form with polyhedral characteristics can be derived:

[0125]

[0126] in, Let n represent the baseline probability of scenario n; then in problem (52) The term is equivalent to

[0127]

[0128] st (55)

[0129]

[0130] Furthermore, step 3.2 also includes: assuming the norm term is a 1 norm, then constraint (52b) is transformed into:

[0131]

[0132] Where t∈[T], i∈[N] b Let ],j,n∈[N]; but

[0133]

[0134] Introducing auxiliary variables will The linear approximation is:

[0135]

[0136] The above equation contains nonlinear terms. By using the McCormick envelope linearization method to perform variable substitution, the nonconvex problem is relaxed into a convex problem:

[0137]

[0138] Simplifying equation (60) yields:

[0139]

[0140] Furthermore, step 3.3 also includes: constraint (30) is a soft constraint in the coupled network, and the following chance constraint is introduced to reduce the conservatism of the final decision:

[0141]

[0142] Its meaning is that in fuzzy sets The probability that the actual hydrogen charging demand at node i satisfies constraint (14) should be no less than 1-σ. i In fuzzy sets The above derivation shows that the chance constraint is equivalent to Z = Z1∪Z2, where

[0143]

[0144]

[0145] Where ξ,η,z n It is an auxiliary variable, ||·|| * It is the dual norm; the chance constraint consists of two inequalities, and equation (62) is transformed into vector form:

[0146]

[0147] Therefore, in the process of opportunity constraint equivalent transformation, the vector function in (63) is replaced by the following equation:

[0148]

[0149] The Z2 constraint in equation (64) is not considered in this scenario because when a(w hy When σ = 0, the chance constraint of the problem will be automatically satisfied; the Z1 constraint is obviously non-convex, and it can be derived that when σ is sufficiently small and σ∈(0,1 / N], Established, replaced by it Then the approximate form of Z1 is:

[0150]

[0151] To reduce the number of auxiliary variables, equation (67) is relaxed to equation (68), which significantly reduces the number of problem variables:

[0152]

[0153] Compared with the prior art, the present invention has at least the following beneficial technical effects:

[0154] 1. This invention aims to solve the "chicken or egg" problem of vehicles and stations in the early stage of hydrogen fuel cell vehicle development. Experiments show that the method proposed in this invention can effectively improve the energy utilization rate of infrastructure, reduce investment and operating costs, effectively avoid energy waste in the early stage of hydrogen vehicle development, and better adapt to the increasing hydrogen demand of hydrogen fuel cell vehicles in the future, providing a valid reference for the future development of hydrogen vehicles and hydrogen infrastructure.

[0155] 2. Through four simulation examples, this invention verifies that the proposed method can reduce grid photovoltaic reduction, improve energy utilization, and effectively alleviate traffic congestion while obtaining more economical and reasonable infrastructure site selection and scale planning decisions.

[0156] 3. By using the improved distributed shaping technique, the probabilities of high-dimensional nonlinear decision-related scenarios introduced by decision-dependency uncertainty are linearized, making the final optimization problem solvable.

[0157] The following will further explain the concept, specific structure, and technical effects of the present invention in conjunction with the accompanying drawings, so as to fully understand the purpose, features, and effects of the present invention. Attached Figure Description

[0158] Figure 1 This is a flowchart of a preferred embodiment of the present invention;

[0159] Figure 2 This is a coupled network framework diagram of a preferred embodiment of the present invention;

[0160] Figure 3 This is a traffic network topology diagram of a preferred embodiment of the present invention;

[0161] Figure 4 This is a power grid topology diagram of a preferred embodiment of the present invention;

[0162] Figure 5 This is a hydrogen pipeline topology diagram of a preferred embodiment of the present invention;

[0163] Figure 6 This is a schematic diagram of the electricity-hydrogen trading price according to a preferred embodiment of the present invention;

[0164] Figure 7 This is photovoltaic power generation prediction data according to a preferred embodiment of the present invention;

[0165] Figure 8 This is a HESI device location planning decision diagram of example 1-4 of a preferred embodiment of the present invention;

[0166] Figure 9 This is a schematic diagram illustrating the impact of the number of selectable HESI planning nodes on investment costs according to a preferred embodiment of the present invention;

[0167] Figure 10 This is a schematic diagram illustrating the impact of the number of selectable HESI planning nodes on unmet hydrogen load in a preferred embodiment of the present invention. Detailed Implementation

[0168] The following description, with reference to the accompanying drawings, illustrates several preferred embodiments of the present invention to make its technical content clearer and easier to understand. The present invention can be embodied in many different forms, and the scope of protection of the present invention is not limited to the embodiments mentioned herein.

[0169] In the accompanying drawings, components with the same structure are indicated by the same numerical designation, and components with similar structures or functions are indicated by similar numerical designations. The dimensions and thicknesses of each component shown in the drawings are arbitrary, and the present invention does not limit the dimensions and thicknesses of each component. To make the illustrations clearer, the thickness of some components has been appropriately exaggerated in the drawings.

[0170] like Figure 1 The diagram shown is a flowchart of a preferred embodiment of the present invention, including the following steps:

[0171] Step 1: Coupling Modeling of the Power Grid, Hydrogen Grid, and Transportation Network. This embodiment fully utilizes the synergistic effect of the coupled networks of the power grid, hydrogen grid, and transportation network. Through the site selection and scale planning of infrastructure, it fully consumes excess renewable energy from the power grid and redistributes traffic flow in the transportation network, achieving joint optimization of the three networks. The specific operations are as follows:

[0172] S1. Construction of the Coupled Network Physical Scenario and Establishment of Optimization Objectives. In the planning of hydrogen-containing hybrid energy supply infrastructure, hydrogen refueling stations, charging stations, electricity-to-gas devices, and hydrogen pipelines are planned based on the charging and hydrogen refueling needs and economic benefits of each node. Their investment decisions are as follows: and w l All are binary decision variables. Assuming the hydrogen refueling capacity of each hydrogen fuel cell vehicle is fixed, the hydrogen load is a discrete variable. The specific architecture of the coupled network of the power grid, hydrogen grid, and transportation network is as follows: Figure 2 As shown in the diagram, the hydrogen grid is coupled to the power grid via P2G (Power-to-Grid) technology, fully utilizing the grid's renewable energy resources. Hydrogen produced by the water electrolysis unit is compressed by a compressor and stored in a hydrogen storage tank (HS). After energy conversion and storage, it is supplied to the HFCV (High-Frequency Hydrogen Capacity) via the HRS (High-Speed ​​Resinerary). When PV (Power Generation) output is in surplus, the portion that the energy storage device (ES) cannot store is converted into hydrogen via P2G to reduce the curtailment rate. When PV output is insufficient, for the hydrogen load that P2G cannot meet, this invention achieves hydrogen purchase from hydrogen sources and flexible hydrogen transfer between hydrogen nodes through planned hydrogen pipelines. Furthermore, this embodiment does not consider returning hydrogen from the hydrogen storage tank to the grid via fuel cell discharge, as multiple electro-hydrogen conversions are extremely inefficient, causing unnecessary resource waste, and fuel cells are expensive; therefore, this approach can be omitted in initial infrastructure planning.

[0173] The optimization objective is to minimize the sum of HESI's annual investment cost and annual average operating cost.

[0174]

[0175]

[0176]

[0177] Among them, c hy c P2G c HS , and c FCS The annual investment costs for HRS, P2G, HS, hydrogen pipelines, and FCS are respectively. and These represent the investment capacities of HRS, P2G, HS, and FCS at node i, respectively. This is a set of nodes. Operating costs include penalties for curtailment of solar power, electricity trading with the main grid, penalties for not meeting charging load requirements, traffic congestion, and penalties for not meeting hydrogen charging load requirements.

[0178] S2. Traffic Network Constraint Construction. Assume a finite-range traffic network containing multiple starting points o and multiple ending points d, with multiple paths p allowing electric vehicles to travel from starting point o to ending point d. Each path p consists of multiple links l. Establish the link traffic at each time step in the traffic network. With path traffic Relational model:

[0179]

[0180]

[0181]

[0182] Furthermore, to ensure that each electric vehicle performs a charging operation once when passing through the transportation network, the following constraint is added, representing the flow on the link at the fast charging station connected to grid node i. With path traffic Relational model:

[0183]

[0184]

[0185]

[0186] According to the US Highway Bureau function, the relationship between the travel time of each vehicle on link l and the link traffic is as follows:

[0187]

[0188] Therefore, the total congestion time on link l is:

[0189]

[0190] Referring to the traffic flow optimization management problem, according to the Wardrop user equilibrium principle, when the optimal decision is achieved in this optimization problem, no vehicle in the traffic network can reduce its total cost by changing its own driving decision. This optimization problem is equivalent to the following Karush-Kuhn-Tucker (KKT) conditions:

[0191]

[0192] Equations (12) and (11) are linearized by applying the Big M method and the piecewise linearization method respectively, and then replaced by equations (13) and (14) respectively.

[0193]

[0194]

[0195] S3. Hydrogen network constraint construction. First, for the HS constraint, the modeling form is similar to ES. Equation (15) is the HS energy balance equation, and Equation (16) restricts the maximum amount of hydrogen injected and released into the HS. The state of charge and recycling conditions of the HS are constrained by Equations (17) and (18).

[0196]

[0197]

[0198]

[0199] H i,0 =H i,T (18)

[0200] The assumption is that a hydrogen pipeline is planned only if a hydrogen supply system (HRS) is planned at both ends of the candidate pipeline. Furthermore, this embodiment stipulates that if a P2G (Power-to-Government) pipeline is planned at a node, an HRS must be planned; however, even if an HRS is planned, a P2G pipeline may not be planned, and the node can purchase hydrogen from a hydrogen source. Therefore, the planning decisions regarding HRS, P2G, and hydrogen pipelines are subject to the following limitations:

[0201]

[0202]

[0203] To make the final optimization problem tractable, a hydrogen pipeline modeling approach based on the concept of regions is applied. It is assumed that each candidate hydrogen node is analogous to a region, and the net flow in each region is balanced, thus yielding the balance constraint (21) and the hydrogen flow rate constraint (22).

[0204]

[0205]

[0206] Meanwhile, the hydrogen pipeline has the capacity to store coils, and the dynamic equations for each hydrogen pipeline storage are (23), (24) and (25), which respectively give the pipeline capacity constraints and cycle conditions.

[0207]

[0208]

[0209]

[0210] The supply and demand balance constraint of the hydrogen network is represented by equation (26), which characterizes the conservation relationship between hydrogen production, hydrogen demand, pipeline transportation, storage, and purchase. Equation (27) constrains the upper and lower limits of hydrogen purchases at each node. In constraint (28), The presence of a hydrogen pipeline connecting the hydrogen source and node i is a binary variable, requiring additional constraints in specific simulation cases based on the specific network construction. Constraint (29) ensures the non-negativity of hydrogen rejection demand. Constraint (30) specifies the upper and lower limits of the total hydrogen demand in HFCV.

[0211]

[0212]

[0213]

[0214]

[0215]

[0216] In equation (30) For the uncertain variables considered in this embodiment, in order to further reduce the decision conservatism of the optimization problem, allow There is a certain probability that constraint (30) will not be satisfied, meaning that a small amount of hydrogen demand may exceed the planned upper limit of this node. Therefore, the soft constraint (30) can be constructed in the form of an opportunity constraint, the specific form and processing method of which will be described in step three.

[0217] S4. Construction of grid constraints. This embodiment uses DC power flow to simulate the operating characteristics of the power grid. Although AC power flow is more accurate, it introduces a large number of nonlinear terms into the optimization problem. Although DC power flow has certain errors, the current scenario is an early planning problem for facilities, and power flow parameters such as voltage magnitude have no significant impact on the model and output. Therefore, it is reasonable to use DC power flow here. Equations (31) and (32) are active power balance and line capacity constraints. Constraint (33) limits the upper and lower limits of the electrical energy purchased by each node from the main grid. Equations (34) and (35) represent the upper and lower limits of P2G capacity and PV curtailment, respectively. Constraint (36) ensures the non-negativity of EV charging demand. Constraint (37) limits the total charging demand of EVs. Approximately (38)-(42) represent the relevant constraints of ES.

[0218]

[0219]

[0220]

[0221]

[0222]

[0223]

[0224]

[0225]

[0226]

[0227]

[0228] E e,0 =E e,T (41)

[0229]

[0230] The actual output of the photovoltaic (PV) power is set as an exogenous uncertainty variable, described by the uncertainty set V. The predicted value of the PV output is... The fluctuation range is The threshold value for uncertain variables is used to adjust the robustness of photovoltaic power output.

[0231]

[0232] S5. Coupling Constraint Construction. Coupling constraints (44) and (45) represent the relationship between vehicle flow and charging demand and hydrogen charging demand, respectively. In addition, the hydrogen production of P2G can be expressed as a steady-state linear function of input power, as shown in equation (46).

[0233]

[0234]

[0235]

[0236] Step 2: Construction of the Decision Dependency-Based Hybrid Optimization Model. To plan HESI more effectively and rationally before the widespread adoption of HFCV, this embodiment constructs a fuzzy set of hydrogen demand (DDU). Based on this, a DDU-DRO planning model is established. The decision dependency mechanism is cleverly used to characterize the uncertainty of hydrogen vehicle refueling demand, effectively addressing the problems of mutual constraints between hydrogen vehicle development and hydrogen infrastructure planning in the early stages of hydrogen vehicle development, resulting in low utilization of hydrogen infrastructure and difficulties in planning. The specific operation is as follows:

[0237] S1. Hydrogen demand decision-making depends on fuzzy set construction. For ease of description, let the uncertain variable hydrogen demand be set as... N bLet N be the total number of nodes, and N be the possible values ​​of the uncertain quantity at each node. s There are 100 scenes. The index is n. In DRO-DDU, it is generally assumed that the DDU set has finite decision-independent supports because assigning a probability of 0 to a particular element is equivalent to excluding it from the supports, achieving a point-to-set mapping where decision dependencies are reflected by a varying probability mass function. For finite supports... The candidate probability distribution in is p n (w hy ), corresponding scenarios And ||p n (w hy )||1=1.

[0238] Furthermore, the Wasserstein metric is defined. as follows:

[0239]

[0240] in, To support the probability distribution set above Ξ, П is the joint probability distribution of u1 and u2, with marginal distributions P1 and P2 respectively. ||·|| is an arbitrary norm, and ||u1-u2|| is the cost of moving a unit mass from distribution P1 to P2. In this embodiment, ||·|| adopts the 1 norm, which has better numerical tractability. Furthermore, the fuzzy set is defined as follows:

[0241]

[0242] This fuzzy set can be viewed as an empirical distribution. A sphere centered at r with radius r, where radius r can explicitly control the conservatism of the decision outcome.

[0243] Consider u t Experience distribution and decision-making hy Related, that is Among them, control u t Probability measure of ∈Ξ It is a decision w hy The function, and the sample space Ξ and w hy Irrelevant. Therefore, the DDU fuzzy set can be written in the following form:

[0244]

[0245] Since this embodiment considers hydrogen demand as discrete support, the equivalent form of the above fuzzy set is:

[0246]

[0247] S2. Construction of the sub-Brussels bar optimization framework. The investment optimization problem constructed in step one can be equivalent to the following sub-Brussels bar optimization.

[0248]

[0249] Where f(·) and h(·) are the investment and operating costs of HESI, respectively. To transform the worst-case expectation of (51) into a manageable form, the reconstructed form of the DDU-DRO model is derived by combining the fuzzy set (50) proposed in S1:

[0250]

[0251]

[0252]

[0253] Step 3: Reconstruction and Solution of the DDU-DRO Problem. To solve the DDU-DRO problem (52), the following transformations are required, including the decision-dependent scenario probabilities. The construction and transformation of norm terms, the linearization of norm terms, and the equivalent transformation of the chance constraints implied in constraint (30).

[0254] S1. Decision-dependent scenario probability construction. An improved distribution shaping technique is proposed, extending and deriving the traditional distribution shaping linearization technique to scenarios where the uncertainty is a non-binary variable. Unlike existing inventions, the planning decision variable in this embodiment is binary, while the uncertainty is a non-binary discrete variable. Step S1 in step two already mentioned that the dimension of u is N. b The possible implementations of u are N s A discrete set of values, i.e., {0,2,...,N} s -1}, with index k, increasing from low to high. The number of scenes is The index is n. For the investment decision at node i, given the parameter vector... and They are respectively The probabilities of the possible values ​​of hydrogen charging demand at node i when taking values ​​of 0 and 1 are obtained from historical data, and there is an internal relationship between them. and The probability of scenario n at node i is

[0255]

[0256] in, Then the decision depends on the probability distribution The probability of scenario n is

[0257]

[0258] The probability expression (54) is a high-dimensional nonlinear function of the decision variables, and the number of scenarios is exponentially related to the number of nodes, resulting in an excessively large formula. Using an improved distributional shaping theory, a set of linear constraints is used to characterize the scenario probabilities related to the decision. By introducing a truncation vector, a linear form with polyhedral characteristics can be derived:

[0259]

[0260] in, Let represent the baseline probability of scenario n. Then, in problem (52)... The term is equivalent to

[0261]

[0262] st (55)

[0263]

[0264] S2, Norm Term Linearization. Assuming the norm term is 1-norm, constraint (52b) is transformed into...

[0265]

[0266] Where t∈[T], i∈[N] b Let ],j,n∈[N]. but

[0267]

[0268] Introducing auxiliary variables will The linear approximation is

[0269]

[0270] The above equation contains nonlinear terms. By using the McCormick envelope linearization method to perform variable substitution, the nonconvex problem is relaxed into a convex problem:

[0271]

[0272] After simplifying equation (60), we get

[0273]

[0274] S3. Opportunity Constraint Reconstruction. Constraint (30) is a soft constraint in the coupled network. The following opportunity constraints can be introduced to reduce the conservatism of the final decision:

[0275]

[0276] Its meaning is that in fuzzy sets The probability that the actual hydrogen charging demand at node i satisfies constraint (14) should be no less than 1-σ. i In fuzzy sets From the above, it can be derived that the chance constraint is equivalent to Z = Z1∪Z2, where

[0277]

[0278]

[0279] Where ξ,η,z n It is an auxiliary variable, ||·|| * It is the dual norm. The chance constraint consists of two inequalities, and equation (62) can be transformed into a vector form.

[0280]

[0281] Therefore, in the process of opportunity constraint equivalent transformation, the vector function in (63) is replaced by the following equation.

[0282]

[0283] The Z2 constraint in equation (64) can be ignored in this scenario because when a(w hy When σ = 0, the chance constraint of the problem will be automatically satisfied. The Z1 constraint is clearly non-convex, and it can be derived that when σ is sufficiently small and σ∈(0,1 / N], It is valid and can be used to replace it. Then the approximate form of Z1 is:

[0284]

[0285] Since some auxiliary variables have been introduced in the solution process, when the amount of historical sample data is large, the excessive computational burden may hinder us from making full use of the data. In order to reduce the number of auxiliary variables, equation (67) can be relaxed to equation (68) to significantly reduce the number of problem variables.

[0286]

[0287] Step 4: Simulation Analysis of a Three-Network Coupled System. In this embodiment, to verify the effectiveness of the proposed planning method, we apply it to a coupled system with a 33-bus power network and a 12-node traffic network. This includes 6 HESI candidate nodes, numbered from H1 to H6, which are connected by 8 candidate hydrogen pipelines. The topology of the three networks is as follows: Figure 3 , Figure 4 , Figure 5As shown in Table 1, the main investment parameters for the HESI device are provided. Candidate nodes H2 and H4 are type 2 devices, while the other nodes are type 1 devices. Figure 6 , Figure 7 It displays the unit purchase price of electricity and hydrogen, as well as forecast data for photovoltaic power generation.

[0288] Table 1. Main Investment Parameters of HESI Unit

[0289] equipment Type 1 Type 2 HRS capacity (kg) 5000 4600 P2G capacity (MW) 50 40 HS Capacity Limits (kg) (Maximum / Minimum) 6000 / 200 5500 / 200 ES Capacity Limit (MW) (Maximum / Minimum) 40 / 7 30 / 5 FCS capacity (MW) 50 40 P2G conversion efficiency 0.79 0.79 P2G electro-hydrogen conversion factor (kg / MW) 28.7 28.7

[0290] This embodiment presents four optimization examples to verify the superiority of the proposed method. Example 1: The hydrogen demand of each node is set as a known parameter, which is an empirical value from historical data; Example 2: The hydrogen demand is set as a robust optimization uncertainty set, with an uncertainty threshold of 6; Example 3: A fuzzy set of hydrogen demand is constructed by applying a DRO model that considers decision-independent uncertainty, where the empirical probability distribution of the Wasserstein metric fuzzy set is assumed to be the Dirac measure; Example 4: The hydrogen demand DRO model with decision-dependent fuzzy sets proposed in this invention is applied. Next, the planning decisions and investment operating costs of HESI are analyzed to see how different modeling methods and the number of candidate nodes in the coupled network affect the results.

[0291] Figure 8 Table 2 shows the corresponding equipment planning and site selection, investment scale, and total investment cost for the four calculation examples. Table 3 shows the operating costs and total planning costs for the four calculation examples; the total planning cost is the sum of HESI's investment cost and average operating cost. From... Figure 8As shown in Table 2, Example 1 and Example 2 have the same facility scale and investment cost, which are the highest among the four examples. In Table 3, Example 2 has a higher total planning cost than Example 1 due to its higher operating cost. This is because it considers the worst-case scenario of photovoltaic power generation, leading to a corresponding increase in operating costs. In Table 2, Examples 3 and 4 use the DRO model, which considers some probabilistic information of uncertain parameters, effectively reducing the conservatism of the final optimization decision. However, in Table 2, Example 3 has the highest operating cost among the four examples because it has less P2G planning, purchases less hydrogen, and forgoes a large amount of hydrogen demand to reduce investment costs, resulting in higher operating costs. Nevertheless, the method in Example 3 can still effectively reduce the conservatism of the decision and lower investment costs in early infrastructure planning. In Example 4, the facility planning scale, investment cost, and operating cost are the lowest among the four examples. Furthermore, each operating cost in Example 4, including traffic congestion costs, hydrogen purchase costs, and unmet hydrogen demand costs, is at a low level. Compared to other examples, this example plans fewer HRS, P2G, FCS, and hydrogen pipelines, and reduces operating costs while ensuring minimal load shedding. Furthermore, only example 4 reduces the number of hydrogen pipeline investments, indicating that local P2G hydrogen production at certain nodes can meet hydrogen load demands. Example 4 applies the DDU-DRO modeling method, considering the impact of planning decisions on the probability distribution of uncertain parameters, thereby appropriately reducing investment while meeting future demand and improving facility utilization.

[0292] Table 2 Investment Scale and Total Investment Cost for Four Examples

[0293] Investment scale and cost Calculation example 1 Calculation example 2 Calculation example 3 Calculation example 4 HRS 6 6 6 5 P2G 6 6 3 3 FCS 4 4 2 2 hydrogen pipeline 5 5 5 4 HESI Investment Cost (M$) 13.29 13.29 8.72 8.13

[0294] Table 3. Operating costs and total planning costs for four case studies.

[0295] Operating costs Calculation example 1 Calculation example 2 Calculation example 3 Calculation example 4 Traffic congestion costs (M$) 0.27 0.88 0.64 0.40 Cost of curtailment (M$) 0 0 0 0 Hydrogen purchase cost (M$) 6.93 7.54 5.08 3.70 Cost of hydrogen withdrawal (M$) 1.73 3.81 11.24 2.52 Total operating costs (M$) 8.93 12.23 16.96 6.62 Total planning cost (M$) 22.22 25.52 25.68 14.75

[0296] In the coupled network, the different planning decisions of HRS and FCS can reallocate traffic flow. Among the four examples above, example 4 has the lowest traffic congestion cost. This shows that the planning decisions obtained by the DDU-DRO modeling method proposed in this embodiment can further optimize traffic network flow allocation and effectively alleviate traffic congestion. In addition, the photovoltaic reduction cost is zero in all four examples, indicating that all four examples prioritize sufficient P2G planning to utilize photovoltaic power generation with zero operating costs. However, only example 4 achieves the same renewable energy consumption as the other examples with the lowest investment and operating costs.

[0297] In addition, this embodiment compares the HESI investment cost and hydrogen supply to hydrogen vehicles under different constraints on the total number of candidate nodes, based on the four examples above. Simulation results Figure 9, Figure 10 The results show that the investment cost and hydrogen demand abandonment cost of Examples 1 and 2 are significantly affected by the total number of candidate nodes, and the trend is almost linear. Examples 3 and 4 show relatively smaller fluctuations, and their investment costs are almost the same under different candidate node number constraints. However, the hydrogen demand abandonment cost of Example 4 decreases significantly with the increase in the number of candidate nodes, while the cost of hydrogen demand abandonment in Example 3 does not decrease significantly. This is because Example 4 considers the increased hydrogen demand due to decision-making resulting from opening more candidate nodes; reducing the number of candidate nodes would weaken the algorithm's performance in adapting to increased demand. Although the hydrogen demand abandonment cost decreases with the increase in the number of candidate nodes for other examples, they incur significant upfront equipment investment costs. This demonstrates that the DDU-DRO modeling method proposed in this embodiment has a significant advantage in adapting to future increases in hydrogen demand.

[0298] This invention addresses the challenges of the early stages of hydrogen vehicle development, where hydrogen vehicle development and hydrogen infrastructure planning are mutually constrained, resulting in low infrastructure utilization and planning difficulties. It develops a planning method for hydrogen-containing hybrid energy supply infrastructure that considers the uncertainty of hydrogen demand decisions. Before hydrogen vehicles become widespread, this method can effectively reduce the conservatism of planning decisions, minimizing investment and operating costs while meeting energy demand, and better adapting to future increases in demand for hydrogen fuel cell vehicles. Furthermore, this method can effectively reduce the curtailment rate of solar power grids and reduce traffic congestion costs, which is of great significance for promoting the widespread adoption of hydrogen fuel cell vehicles. The main contributions of this invention are as follows:

[0299] To improve the operational flexibility of the energy system, a three-network coupling system of hydrogen, transportation, and power grid is established. The hydrogen network is coupled with the power grid not only through P2G (peer-to-grid) but also coordinates with the transportation network by supplying hydrogen to the high-frequency magnetic resonance (HFCV) via a high-speed rail system (HRS). The proposed HESI planning model utilizes P2G technology to enhance renewable energy absorption and reallocates transportation network traffic through infrastructure site selection planning to alleviate traffic congestion.

[0300] A fuzzy set of uncertainty in hydrogen demand decision-making is constructed, and the actual probability distribution of hydrogen demand is set as a Wasserstein sphere centered on an empirical distribution that depends on the site selection and scale planning decisions of the HRS. Furthermore, a two-stage decision-dependent distributed robust programming model is constructed, and an improved distribution shaping method is introduced to handle the high-dimensional nonlinear function caused by the decision-dependent mechanism. Robust equivalence transformation and chance constraint handling methods are applied to derive the solvable form of the DRO problem.

[0301] Simulation results demonstrate that the proposed method can achieve more economical and rational HESI site selection and scale planning decisions in the early stages of hydrogen vehicle development, reducing photovoltaic depletion, improving energy efficiency, and effectively alleviating traffic congestion. Furthermore, simulation results suggest that the DDU-DRO method can better adapt to the increasing hydrogen refueling demand of future hydrogen vehicles compared to other traditional planning methods.

[0302] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A planning method for hydrogen-containing hybrid energy supply infrastructure, characterized in that, The method includes the following steps: Step 1: Model the coupling of the power grid, hydrogen grid, and transportation network; Step 2: Constructing the decision-dependent sub-bar optimization model; Step 3: Reconstruction and solution method of the Bruker optimization problem; Step 4: Simulation analysis of a three-network coupled system; Step 1 further includes: Step 1.1: Constructing the physical scenario of the coupled network and establishing optimization objectives; Step 1.2: Construction of transportation network constraints; Step 1.3: Construction of hydrogen network constraints; Step 1.4: Constructing power grid constraints; Step 1.5: Constructing coupling constraints; Step 2 also includes: Step 2.1: Hydrogen demand decision depends on fuzzy set construction; Step 2.2: Construct the Brussels Bar Optimization Framework; Step 3 also includes: Step 3.1: Constructing the probability of decision-dependent scenarios; Step 3.2: Linearization of norm terms; Step 3.3: Opportunity Constraint Restructuring; Step 1.1 further includes: in the planning of hydrogen-containing hybrid energy supply infrastructure, planning for hydrogen refueling stations, charging stations, electricity-to-gas devices, and hydrogen pipelines is carried out based on the charging and hydrogen refueling needs and economic benefits of the nodes, and the investment decisions are as follows: , , and All variables are binary decision variables; assuming a fixed hydrogen refueling capacity for each hydrogen fuel cell vehicle, the hydrogen load is a discrete variable; the hydrogen grid is coupled to the power grid via a power-to-gas (P2G) device, making full use of the grid's renewable energy. Hydrogen produced by the water electrolysis unit is compressed by a compressor and stored in a hydrogen storage tank (HS). After energy conversion and storage, it is supplied to hydrogen fuel cell vehicles (HFCVs) through hydrogen refueling stations (HRS); when PV output is in surplus, the portion that the energy storage device (ES) cannot store is converted into hydrogen via P2G to reduce the curtailment rate; when PV output is insufficient, the hydrogen load that P2G cannot meet is addressed. The optimization objective is to minimize the sum of HESI's annual investment cost and annual average operating cost, as follows: (1) (2) (3) in, , , , and The annual investment costs for HRS, P2G, HS, hydrogen pipelines, and FCS are respectively. , , and These represent the investment capacities of HRS, P2G, HS, and FCS at node i, respectively. For a set of nodes; Operating costs include penalties for curtailment of solar power, electricity trading with the main grid, penalties for not meeting charging load requirements, traffic congestion, and penalties for not meeting hydrogen charging load requirements. Step 1.2 further includes: assuming a finite-range transportation network containing multiple starting points. o and multiple endpoints d And there are multiple paths p This allows electric vehicles to start from the starting point o To the finish line d Each path p All consist of multiple links l Composition; Establishing the link traffic at each moment in the transportation network With path traffic Relational model: (4) ， (5) (6) To ensure that every electric vehicle undergoes a charging operation when passing through this transportation network, the following constraint is added, indicating that at the grid node... i Traffic on the link at the connected fast charging station With path traffic Relational model: (7) (8) (9) According to the US Road Transport Authority function, each vehicle is on the link l The relationship between travel time and link traffic is as follows: (10) link l The total congestion time is: (11) Referring to the traffic flow optimization management problem, according to the Wardrop user equilibrium principle, when the optimal decision is achieved in this optimization problem, no vehicle in the traffic network can reduce its total cost by changing its own driving decision; this optimization problem is equivalent to the following Karush-Kuhn-Tucker (KKT) conditions: (12) Equations (12) and (11) are linearized using the Big M method and piecewise linearization method, respectively, and then replaced with equations (13) and (14), respectively: (13) (14) Step 1.3 further includes: for HS constraints, equation (15) is the HS energy balance equation, equation (16) limits the maximum hydrogen injection and release of HS, and the state of charge and recycling conditions of HS are constrained by equations (17) and (18): (15) (16) (17) (18) The following limitations exist in planning decisions regarding HRS, P2G, and hydrogen pipelines: (19) (20) To make the final optimization problem easier to handle, a hydrogen pipeline modeling method based on the concept of regions is applied; assuming that each candidate hydrogen node is analogous to a region and that the net flow of each region is balanced, the balance constraint (21) and the hydrogen flow rate limit (22) are obtained: (21) (22) Meanwhile, the hydrogen pipelines have the capacity for coil storage, and the storage dynamic equations for each hydrogen pipeline are (23), (24) and (25), which respectively give the pipeline capacity constraints and cyclic conditions: (23) (24) (25) The supply and demand balance constraint of the hydrogen network is represented by equation (26), which characterizes the conservation relationship of hydrogen production, hydrogen demand, pipeline transportation, storage, and purchase: (26) Equation (27) constrains the upper and lower limits of hydrogen purchases at each node: (27) In constraint (28), Indicates whether there is a connection between the hydrogen source and the node. i The hydrogen pipeline is a binary variable, requiring additional constraints to be added in specific simulation cases based on the specific network construction: (28) Constraint (29) ensures the nonnegativity of hydrogen abandonment demand: (29) Constraint (30) specifies the upper and lower limits of the total hydrogen demand for the HFCV: (30) In equation (30) It is an uncertain variable; Step 1.4 further includes: Equations (31) and (32) are for active power balance and line capacity constraints: (31) (32) Constraint (33) limits the upper and lower limits of the amount of electricity each node can purchase from the main grid: (33) Equations (34) and (35) represent the upper and lower limits of P2G capacity and PV waste, respectively: (34) (35) Constraint (36) ensures the nonnegativity of the charging demand for abandoned EVs: (36) Constraint (37) limits the total charging demand of EVs: (37) Constraints (38)-(42) represent the relevant constraints of ES: (38) (39) (40) (41) (42) The actual output of the photovoltaic (PV) power is set as an exogenous uncertain variable, described by the uncertainty set V; where the predicted value of the PV output is... The fluctuation range is , The threshold value for the uncertain variable is used to adjust the robustness of photovoltaic power output. (43) Step 1.5 further includes: coupling constraints (44) and (45) representing the relationship between vehicle flow and charging demand and hydrogen refueling demand, respectively: (44) (45) The hydrogen production of P2G is expressed as a steady-state linear function of the input power, as shown in equation (46): (46) Step 2.1 further includes: setting the uncertain variable hydrogen demand as... , Let be the total number of nodes, and the possible values ​​of the uncertainties at each node be: There are 100 scenes. The index is n In DRO-DDU, it is generally assumed that the DDU set has finite decision-independent supports because assigning a probability of 0 to a particular element is equivalent to excluding it from the supports, achieving a point-to-set mapping, where decision dependence is reflected by a varying probability quality function; for finite supports , The candidate probability distribution in is Corresponding scenarios ,and ; Define Wasserstein metric as follows: (47) in, To support The probability distribution set, for and The joint probability distributions are P1 and P2, respectively; For any norm, It is the cost of moving a unit mass from distribution P1 to P2; Fuzzy sets are defined as follows: (48) This fuzzy set can be viewed as an empirical distribution. Centered on, with r A sphere with radius , radius r It can clearly control the conservatism of decision-making outcomes; consider Experience distribution and decision-making Related, that is Among them, control probability measure It is a decision The function, and the sample space and Irrelevant; DDU fuzzy sets are written in the following form: (49) Since the hydrogen demand is discretely supported, the equivalent form of the above fuzzy set is: (50) Step 2.2 further includes: the investment optimization problem constructed in step one is equivalent to the following sub-Bluerg optimization. s.t. (51) in, and These represent the investment and operating costs of HESI, respectively; combining the fuzzy set (50) proposed in step 2.1, the reconstructed form of the DDU-DRO model is derived: (52a) s.t. (52b) (52c) (52d,52e)。 2. The planning method for hydrogen-containing hybrid energy supply infrastructure as described in claim 1, characterized in that, Step 3.1 further includes: as mentioned in step 2.1 u The dimension is , u The possible implementation is A discrete value, i.e. The index is k The number of scenarios increases from low to high. The index is n ; For nodes i Investment decisions at a given location; given parameter vector and , respectively When taking 0 and 1, the node i The probability of the hydrogen charging demand taking a given value is obtained from historical data, and there is an internal relationship between them. and Then the node i The scene n The probability is: (53) in, , Then the decision depends on the probability distribution. Mid-scene n The probability is: (54) The probability expression (54) is a high-dimensional nonlinear function of the decision variable, and the number of scenarios is exponentially related to the number of nodes. Using an improved distribution shaping theory, a set of linear constraints is used to characterize the scenario probabilities related to the decision. By introducing a truncation vector, a linear form with polyhedral characteristics can be derived: (55) in, , , indicating a scene n The baseline probability; then in problem (52) The term is equivalent to st (55) (56)。 3. The planning method for hydrogen-containing hybrid energy supply infrastructure as described in claim 2, characterized in that, Step 3.2 further includes: assuming the norm term is a 1 norm, then constraint (52b) is transformed into: (57) in, ;make ,but (58) Introducing auxiliary variables will The linear approximation is: (59) The above equation contains nonlinear terms. By using the McCormick envelope linearization method to perform variable substitution, the nonconvex problem is relaxed into a convex problem: (60) Simplifying equation (60) yields: (61)。 4. The planning method for hydrogen-containing hybrid energy supply infrastructure as described in claim 3, characterized in that, Step 3.3 further includes: Constraint (30) is a soft constraint in the coupled network, and the following chance constraints are introduced to reduce the conservatism of the final decision: (62) Its meaning is that in fuzzy sets Next node i The probability that the actual hydrogen charging demand satisfies constraint (14) should be no less than In fuzzy sets Above, the opportunity constraint is derived to be equivalent to ,in (63) (64) in, It is an auxiliary variable. It is the dual norm; the chance constraint consists of two inequalities, and equation (62) is transformed into vector form: (65) Therefore, in the process of opportunity constraint equivalent transformation, the vector function in (63) is replaced by the following equation: (66) Equation (64) Constraints are not considered in this scenario because when The chance constraints of this problem will be automatically satisfied; the Z1 constraint is clearly non-convex, as can be derived. Small enough and hour, Established, replaced by it Then the approximate form of Z1 is: (67) To reduce the number of auxiliary variables, equation (67) is relaxed to equation (68), which significantly reduces the number of problem variables: (68)。

Images