Hydrogen refueling station planning and operation optimization method considering traffic-service two-layer network robustness

Abstract

The application discloses a hydrogen refueling station planning operation optimization method considering traffic-service double-layer network resilience, comprising the following steps: constructing an upper layer site selection and capacity planning model with the maximum daily profit of a hydrogen refueling station system and the minimum comprehensive vulnerability of a double-layer network as upper layer targets; constructing a lower layer operation model with the maximum user comprehensive satisfaction as a lower layer target, thereby obtaining a hydrogen refueling station planning operation double-layer optimization model; and adopting a distribution robust optimization method based on a Wasserstein distance in combination with a MAPPO algorithm to solve the hydrogen refueling station planning operation double-layer optimization model, so as to obtain an optimal site selection and capacity scheme and an optimal demand flow distribution scheme. The application can strengthen the resilience of the double-layer network, optimize network layout and risk resistance, and realize the collaborative optimization of economy, resilience and service quality.

Description

Technical Field

[0001] This invention relates to the field of hydrogen refueling station planning technology, and in particular to a method for optimizing the planning and operation of hydrogen refueling stations that takes into account the resilience of the traffic-service two-layer network. Background Technology

[0002] Hydrogen refueling stations are core service facilities connecting hydrogen energy supply and demand, and building a hydrogen refueling station network is a fundamental support for promoting hydrogen energy development and constructing a new energy system. However, due to the low penetration rate and unreasonable layout of hydrogen refueling stations, the existing hydrogen refueling station network lacks sufficient resilience and robustness in the face of demand fluctuations and node failures, which seriously restricts the large-scale application and development of the hydrogen energy industry. The layout of the hydrogen refueling station network cannot effectively connect the supply and demand of hydrogen, and there is an urgent need to strengthen the resilience and optimize the layout of the hydrogen refueling station network.

[0003] System resilience, which enables systems to recover from extreme events or disasters and further improve, has become an important topic in the field of complex networks. Currently, the definition of resilience is diverse, with academia primarily defining it from the perspectives of capability, process, and outcome.

[0004] Current research on hydrogen refueling station site selection mainly includes two aspects: hydrogen demand based on network nodes and traffic demand based on road networks, as well as comprehensive mathematical site selection models. Starting from network node demand, the main approaches include the P-median model and the set-covering model.

[0005] Deep reinforcement learning (DRL) can learn optimal strategies through continuous interaction with the environment, providing a new technical approach to solving facility location problems. Currently, its applications in location problems mainly include classical static location, dynamic multi-period location, joint location-path optimization, and electric vehicle charging station location. Existing research shows that DRL can achieve efficient and robust decision-making in various location scenarios, demonstrating broad application prospects.

[0006] Existing research has laid an important theoretical foundation for the layout of hydrogen refueling station networks, but the following problems still exist: existing research focuses on economic analysis and lacks a network resilience perspective; it focuses on the site selection of single sites and rarely involves two-layer network layout; it is based on deterministic assumptions and is difficult to characterize demand uncertainty; traditional methods are difficult to efficiently solve large-scale spatiotemporal coupled decisions.

[0007] Therefore, a hydrogen refueling station planning and operation optimization method that takes into account the resilience of the traffic-service two-layer network is needed to solve the problems existing in the above schemes. Summary of the Invention

[0008] Therefore, this invention provides a hydrogen refueling station planning and operation optimization method that takes into account the resilience of the traffic-service two-layer network, in order to solve or at least alleviate the problems mentioned above.

[0009] According to one aspect of the present invention, a method for optimizing the planning and operation of hydrogen refueling stations considering the resilience of a two-layer transportation-service network is provided. This method is executed in a computing device. The two-layer network comprises a service network and a transportation network. The service network is adapted to consist of hydrogen refueling stations supplying hydrogen and demand points generating hydrogen demand. The transportation network is adapted to consist of hydrogen refueling stations, demand points, hydrogen fuel cell vehicles, and roads. All hydrogen refueling stations within the two-layer network form a hydrogen refueling station system. The method includes: constructing an upper-layer site selection and capacity planning model with the upper-layer objective of maximizing the daily profit of the hydrogen refueling station system and minimizing the overall vulnerability of the two-layer network. The daily profit of the hydrogen refueling station system is the difference between the total daily hydrogen sales revenue and the daily construction and operation costs of the hydrogen refueling station system. The daily construction and operation costs include the daily construction investment costs and daily operating costs. The overall vulnerability is adapted to be determined based on the average vulnerability of nodes, the average vulnerability of edges, and the network... Global vulnerability is determined; a lower-level operation model is constructed with maximizing overall user satisfaction as the lower-level objective. Overall user satisfaction is determined based on time satisfaction and economic satisfaction values. The time satisfaction value characterizes the hydrogen refueling service quality of the two-layer network, and the economic satisfaction value characterizes the economic cost of hydrogen refueling for users. Based on the upper-level site selection and capacity planning model and the lower-level operation model, a two-layer optimization model for hydrogen refueling station planning and operation is obtained. The bibliometric optimization method based on Wasserstein distance combined with the MAPPO algorithm is used to solve the two-layer optimization model for hydrogen refueling station planning and operation, obtaining the optimal site selection and capacity allocation scheme and the optimal demand flow allocation scheme. The optimal site selection and capacity allocation scheme indicates whether each candidate point is selected as a hydrogen refueling station, the construction capacity of each hydrogen refueling station, and whether each hydrogen refueling station serves each demand point. The optimal demand flow allocation scheme indicates the amount of hydrogen each hydrogen refueling station meets for each demand point at each time period.

[0010] Optionally, in the hydrogen refueling station planning and operation optimization method of the present invention that takes into account the resilience of the traffic-service two-layer network, the total daily hydrogen sales revenue is suitable for determination based on the hydrogen demand met by each hydrogen refueling station in each time period and the hydrogen retail price; the daily construction investment cost of the hydrogen refueling station system is suitable for determination based on the land investment cost of all hydrogen refueling stations, the equipment investment cost of all hydrogen refueling stations, and the number of operating days per year; the daily operating cost of the hydrogen refueling station system includes the total daily hydrogen procurement cost, the total daily equipment operation and maintenance cost, and the total daily labor cost.

[0011] Optionally, in the hydrogen refueling station planning and operation optimization method of the present invention that takes into account the resilience of the traffic-service two-layer network, the land investment cost of all hydrogen refueling stations is equal to the sum of the construction land costs of all hydrogen refueling stations, and the construction land cost of each hydrogen refueling station is related to the land price and area of ​​the location of the hydrogen refueling station; the equipment investment cost of all hydrogen refueling stations includes the purchase cost of compressors, hydrogen storage tanks, and hydrogen refueling machines of all hydrogen refueling stations; the total daily hydrogen procurement cost is suitable for determination based on the hydrogen demand met by each hydrogen refueling station in each time period and the unit price of hydrogen procurement; the total daily equipment operation and maintenance cost is the sum of the daily equipment operation and maintenance costs of all hydrogen refueling stations; the total daily labor cost is the sum of the daily labor costs of all hydrogen refueling stations.

[0012] Optionally, in the hydrogen refueling station planning and operation optimization method considering the resilience of the transportation-service two-layer network according to the present invention, an upper-layer site selection and capacity planning model is constructed with the upper-layer objective of maximizing the daily profit of the hydrogen refueling station system and minimizing the overall vulnerability of the two-layer network. This includes: constructing an upper-layer first objective function with the upper-layer objective of maximizing the daily profit of the hydrogen refueling station system; constructing an upper-layer second objective function with the upper-layer objective of minimizing the overall vulnerability of the two-layer network, based on the average vulnerability of nodes, average vulnerability of edges, and global vulnerability of the network using the ideal point method; and constructing an upper-layer site selection and capacity planning model based on the first objective function and the second objective function.

[0013] Optionally, in the hydrogen refueling station planning and operation optimization method considering the resilience of the transportation-service two-layer network according to the present invention, an upper-layer site selection and capacity planning model is constructed based on the first objective function and the second objective function, including: constructing an upper-layer site selection and capacity planning model based on the first objective function, the second objective function, and various upper-layer constraints; the various upper-layer constraints include: site coverage constraints, demand satisfaction constraints, logical association constraints, vulnerability threshold constraints, decision variable domain constraints, capacity standardization constraints, capacity range constraints, land area constraints, two-layer network scale constraints, and hydrogen refueling station configuration constraints.

[0014] Optionally, in the hydrogen refueling station planning and operation optimization method considering the resilience of the traffic-service two-layer network according to the present invention, the average vulnerability of nodes is suitable for determination based on the vulnerability of each hydrogen refueling station, wherein the vulnerability of each hydrogen refueling station is suitable for obtaining by normalizing the weighted clustering coefficient of the hydrogen refueling station; the average vulnerability of edges is suitable for determination based on the vulnerability of the path between each demand point and each hydrogen refueling station, and is related to the probability of interruption of the path between each demand point and each hydrogen refueling station and the change in the average path length of the network caused by the path interruption; the global vulnerability of the network is suitable for obtaining by normalizing the global network efficiency.

[0015] Optionally, in the hydrogen refueling station planning and operation optimization method considering the resilience of the traffic-service two-layer network according to the present invention, a lower-level operation model is constructed with the maximum overall user satisfaction as the lower-level objective, including: constructing a lower-level objective function with the maximum overall user satisfaction as the lower-level objective, and constructing a lower-level operation model based on the lower-level objective function; wherein, the overall user satisfaction is suitable for obtaining using the TOPSIS method based on time satisfaction value and economic satisfaction value.

[0016] Optionally, in the hydrogen refueling station planning and operation optimization method considering the resilience of the traffic-service two-layer network according to the present invention, constructing a lower-level operation model based on the lower-level objective function includes: constructing a lower-level operation model based on the lower-level objective function and various lower-level constraints; the various lower-level constraints include: service relationship coupling constraints, demand satisfaction upper limit constraints, instantaneous service rate constraints of hydrogen refueling stations, causal constraints of hydrogen fuel cell vehicle driving time, daily capacity constraints of hydrogen refueling stations, and load balancing constraints of hydrogen refueling stations.

[0017] Optionally, in the hydrogen refueling station planning and operation optimization method considering the resilience of the traffic-service two-layer network according to the present invention, a biblical bar optimization method based on Wasserstein distance combined with the MAPPO algorithm is used to solve the two-layer optimization model for hydrogen refueling station planning and operation, obtaining the optimal site selection and capacity allocation scheme and the optimal demand flow allocation scheme. This includes: obtaining model input parameters, which include parameters of multiple candidate points and multiple demand points in the two-layer network, and historical time-series hydrogen demand data; constructing a historical hydrogen demand empirical distribution based on the historical time-series hydrogen demand data; using the biblical bar optimization method based on Wasserstein distance to construct a distributed fuzzy set centered on the historical hydrogen demand empirical distribution to quantify the uncertainty of hydrogen demand distribution; and then applying the hydrogen refueling station planning... The bilayer optimization model is transformed into a bilayer optimization model with embedded uncertainties in hydrogen demand distribution. Maximizing daily profit and minimizing overall vulnerability of the upper-layer site selection and capacity planning model are taken as primary and secondary objectives, respectively. A spectral clustering algorithm is used to aggregate the multiple demand points into four demand alliances, establishing agents for each alliance and configuring observation and action spaces for each agent to construct a multi-agent reinforcement learning environment. Based on this environment, each agent executes joint site selection and flow allocation actions based on its corresponding observation space. The value of these joint actions is centrally evaluated through a global value network, and the policy network is updated through a policy pruning mechanism to iteratively solve the bilayer optimization model, obtaining the optimal site selection and capacity planning scheme and the optimal demand flow allocation scheme.

[0018] According to one aspect of the present invention, 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 hydrogen refueling station planning and operation optimization method considering the resilience of a two-tier transportation-service network as described above.

[0019] According to one aspect of the present invention, a computer program product is provided, comprising a computer program / instructions, wherein the computer program / instructions, when executed by a processor, implement the method as described above.

[0020] According to one aspect of the present invention, a readable storage medium storing program instructions is provided, which, when read and executed by a computing device, causes the computing device to perform the hydrogen refueling station planning and operation optimization method as described above, taking into account the resilience of the traffic-service two-layer network.

[0021] According to the technical solution of this invention, a method for optimizing the planning and operation of hydrogen refueling stations considering the resilience of a two-layer transportation-service network is provided. Based on establishing a two-layer network vulnerability assessment model, a two-layer optimization model for hydrogen refueling station planning and operation is constructed, with the upper-layer objective of maximizing daily profit and minimizing overall vulnerability, and the lower-layer objective of maximizing overall user satisfaction. A distributed bar optimization method is introduced to characterize the uncertainty of demand distribution, and the MAPPO algorithm is combined to achieve efficient solution. This strengthens the resilience of the hydrogen refueling station network, optimizes the network layout and risk resistance, and enables the two-layer network layout to effectively connect hydrogen supply and demand, achieving synergistic optimization of economy, resilience, and service quality. Furthermore, based on the two-layer network architecture, the upper-layer service network issues scheduling instructions, and the lower-layer transportation network provides feedback on user satisfaction, forming a "planning-execution-feedback" closed loop, which can effectively avoid cascading risks caused by single-point failures.

[0022] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0023] To achieve the foregoing and related objectives, certain illustrative aspects of the invention are described in conjunction with the following description and accompanying drawings. These aspects indicate various ways in which the principles of the invention can 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.

[0024] Figure 1A schematic diagram of a two-layer network 100 for a hydrogen refueling station according to an embodiment of the present invention is shown. Figure 2 A schematic diagram of a computing device 200 provided according to an embodiment of the present invention is shown; Figure 3 A flowchart illustrating a hydrogen refueling station planning and operation optimization method 300 considering the resilience of a two-layer transportation-service network, according to an embodiment of the present invention, is shown. Figure 4 A schematic diagram of the distribution of demand points according to some embodiments of the present invention is shown; Figure 5 A schematic diagram showing the distribution of candidate hydrogen refueling station sites according to some embodiments of the present invention is provided. Figure 6 The diagram illustrates the hydrogenation probability distribution under various scenarios according to some embodiments of the present invention; Figure 7 A schematic diagram showing the location distribution of hydrogen refueling stations according to some embodiments of the present invention is shown; Figure 8 A schematic diagram illustrating the service relationship between hydrogen refueling stations and demand points according to some embodiments of the present invention is shown. Figure 9 A schematic diagram of the total service traffic distribution for different time periods is shown in some embodiments of the present invention. Detailed Implementation

[0025] 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.

[0026] To address the shortcomings of existing hydrogen refueling station network layout schemes, which often focus on economic analysis and lack a network resilience perspective, and primarily address single-site site selection while rarely considering two-layer network layouts, making it difficult to characterize demand uncertainty, this invention proposes a hydrogen refueling station planning and operation optimization method that considers the resilience of a two-layer transportation-service network. Based on a two-layer network vulnerability assessment model, a two-layer optimization model for hydrogen refueling station planning and operation is constructed, with the upper-layer objective of maximizing daily profit and minimizing overall vulnerability, and the lower-layer objective of maximizing overall user satisfaction. The method introduces the Bruker bar optimization technique to characterize demand distribution uncertainty and combines it with the MAPPO algorithm for efficient solution. This enhances the resilience of the hydrogen refueling station network, optimizes network layout and risk resistance, and achieves synergistic optimization of economic efficiency, resilience, and service quality.

[0027] Figure 1A schematic diagram of a traffic-service two-layer network 100 constructed according to an embodiment of the present invention is shown.

[0028] like Figure 1 As shown, the transportation-service two-layer network 100 comprises two layers: a service network and a transportation network. This transportation-service two-layer network 100 is abstracted as a two-layer mapping model composed of a heterogeneous set of nodes. The upper layer is the service network, consisting of hydrogen refueling stations supplying hydrogen (meeting hydrogen demand) and demand points generating hydrogen demand. The connections between nodes (connections between hydrogen refueling stations and demand points) represent the service relationship between hydrogen refueling stations and demand points, reflecting the supply-demand matching logic. The lower layer is the transportation network, composed of hydrogen refueling stations, demand points, hydrogen fuel cell vehicles, and roads. Hydrogen fuel cell vehicles travel along roads and enter hydrogen refueling stations to complete refueling, depicting the actual movement and energy replenishment process in physical space. The two layers of networks achieve bidirectional interaction through coupling: the upper service network issues scheduling instructions to the lower transportation network, constraining the route selection and refueling sequence of hydrogen fuel cell vehicles; the lower transportation network feeds back user satisfaction to the upper service network, providing data support for iterative correction of the upper-layer planning scheme. Thus, the supply and demand matching logic in the service network and the physical movement logic in the transportation network are organically combined to form a closed-loop decision-making system of "planning-execution-feedback".

[0029] In the transportation-service two-layer network 100 constructed according to embodiments of the present invention, each component assumes a differentiated role based on its level: In the upper-layer service network, hydrogen refueling stations are the core nodes for hydrogen fuel replenishment, responsible for high-pressure or cryogenic hydrogen storage and distributing it to hydrogen fuel cell vehicles on demand; demand points are geographical aggregations of hydrogen energy consumption, typically located in logistics bases, wholesale markets, railway freight yards, and along major passenger and freight corridors, representing the refueling demand of hydrogen fuel cell vehicles within the region. In the lower-layer transportation network, hydrogen fuel cell vehicles, as mobile consumption carriers, establish dynamic connections with hydrogen refueling stations through driving on actual roads, transforming the service demands of the upper layer into executable physical flows at the lower layer; roads constitute the basic support for the passage of hydrogen fuel cell vehicles. The two-layer network has the following characteristics: Structurally, the upper-layer service network and the lower-layer transportation network are independent of each other, and each forms its internal topology through service relationships or road connections; in terms of association, the two layers bind service logic and physical location through coupling; in terms of function, the supply and demand matching determined by the upper-layer service network is transmitted to the lower-layer transportation network through coupling, driving the actual hydrogen refueling behavior of hydrogen fuel cell vehicles, forming a complete closed loop of "planning-execution-feedback".

[0030] In this embodiment of the invention, a two-layer network vulnerability assessment model is pre-established. According to the two-layer network vulnerability assessment model, the overall vulnerability of the traffic-service two-layer network is related to the average vulnerability of nodes, the average vulnerability of edges, and the global vulnerability of the network. The average vulnerability of nodes, the average vulnerability of edges, and the global vulnerability of the network can be represented by the normalized weighted clustering coefficient, the change in the average path length of the network caused by path interruption, and the global network efficiency, respectively.

[0031] The vulnerability of each hydrogen refueling station can be obtained by normalizing the weighted clustering coefficient of the hydrogen refueling station. The weighted clustering coefficient of the hydrogen refueling station is related to the original weight of the hydrogen refueling station.

[0032] The original weights of hydrogen refueling stations are as follows: (1) In the formula, The original weights of hydrogen refueling station j; Let be the weight of the path between demand point i and hydrogen refueling station j, and be the reciprocal of the distance.

[0033] The weighted clustering coefficients for hydrogen refueling stations are as follows: (2) In the formula, Let be the weighted clustering coefficient of hydrogen refueling station j. The degree of hydrogen refueling station j.

[0034] Specifically, by standardizing the weighted clustering coefficients of hydrogen refueling station j and then subtracting them from 1, the vulnerability of hydrogen refueling station j can be obtained: (3) The average vulnerability of nodes in the two-layer network can be calculated based on the vulnerability of each hydrogen refueling station in the network, as shown in the following formula: (4) The probability of a path being interrupted is negatively correlated with the actual road network distance and the number of alternative paths, as shown in the following formula: (5) In the formula, Let be the probability of path interruption between demand point i and hydrogen refueling station j; The number of alternative paths (road redundancy) between demand point i and hydrogen refueling station j. The actual road network distance between demand point i and hydrogen refueling station j.

[0035] The change in average network path length caused by path interruption is used to characterize the impact of path interruption on a two-layer network. Path vulnerability is related to the probability of path interruption and the change in average network path length caused by path interruption. Specifically, the vulnerability of the path between demand point i and hydrogen refueling station j is related to the probability of path interruption between demand point i and hydrogen refueling station j and the change in average network path length caused by path interruption, as shown in the following equation: (6) In the formula, This represents the change in the average network path length caused by the interruption of the path between demand point i and hydrogen refueling station j.

[0036] The average vulnerability of the edges in the two-layer network can be calculated based on the vulnerability of the paths between each demand point and each hydrogen refueling station, as shown in the following formula: (7) Global network efficiency As shown in the following formula: (8) In the formula, This represents the number of demand points in the set.

[0037] The global vulnerability of a two-layer network can be obtained by normalizing the global network efficiency, as shown in the following equation: (9) In the formula, E is a network-wide vulnerability in two-layer networks. max This represents the maximum global network efficiency (i.e., 1).

[0038] In embodiments of the present invention, a computing device may be configured to execute a hydrogen refueling station planning and operation optimization method 300 that considers the resilience of a two-layer transportation-service network. The hydrogen refueling station planning and operation optimization method 300 of the present invention, which considers the resilience of a two-layer transportation-service network, will be described below.

[0039] The following describes a computing device 200 provided by an embodiment of the present invention.

[0040] Figure 2 A schematic diagram of a computing device 200 according to an embodiment of the present invention is shown. Figure 2As shown, in a basic configuration, computing device 200 includes at least one processing unit 202 and system memory 204. According to one aspect, depending on the configuration and type of the computing device, the processing unit 202 may be implemented as a processor. System memory 204 includes, but is not limited to, volatile memory (e.g., random access memory), non-volatile memory (e.g., read-only memory), flash memory, or any combination of such memories. According to one aspect, system memory 204 includes an operating system 205.

[0041] According to one aspect, operating system 205 is, for example, suitable for controlling the operation of computing device 200. Furthermore, examples are practiced in conjunction with graphics libraries, other operating systems, or any other applications, and are not limited to any particular application or system. Figure 2 The basic configuration is illustrated by the components within the dashed lines. According to one aspect, the computing device 200 has additional features or functions. For example, according to one aspect, the computing device 200 includes additional data storage devices (removable and / or non-removable), such as disks, optical discs, or magnetic tapes. This additional storage... Figure 2 The middle part is shown by removable storage device 209 and non-removable storage device 210.

[0042] As stated above, according to one aspect, program module 203 is stored in system memory 204. According to one aspect, program module 203 may include one or more applications. The present invention does not limit the type of application; for example, applications may include: email and contact applications, word processing applications, spreadsheet applications, database applications, slideshow applications, drawing or computer-aided applications, web browser applications, etc.

[0043] In an embodiment of the present invention, program module 203 includes multiple program instructions for executing the hydrogen refueling station planning and operation optimization method 300 of the present invention, which takes into account the resilience of the traffic-service two-layer network.

[0044] According to one aspect, examples can be practiced on circuits including discrete electronic components, packaged or integrated electronic chips containing logic gates, circuits utilizing microprocessors, or on a single chip containing electronic components or a microprocessor. For example, it can be practiced via wherein... Figure 2Each or many of the components shown can be implemented as an example by integrating a System-on-a-Chip (SOC) on a single integrated circuit. According to one aspect, such an SOC device may include one or more processing units, graphics units, communication units, system virtualization units, and various application functions, all integrated (or “burned in”) as a single integrated circuit onto a chip substrate. When operated via the SOC, the functions described in this invention can be operated via dedicated logic integrated on a single integrated circuit (chip) with other components of the computing device 200. Embodiments of the invention can also be implemented using other techniques capable of performing logical operations (e.g., AND, OR, and NOT), including but not limited to mechanical, optical, fluid, and quantum technologies. Additionally, embodiments of the invention can be implemented within a general-purpose computer or in any other circuit or system.

[0045] According to one aspect, computing device 200 may also have one or more input devices 212, such as a keyboard, mouse, pen, voice input device, touch input device, etc. It may also include output devices 214, such as a display, speaker, printer, etc. The foregoing devices are examples and other devices may also be used. Computing device 200 may include one or more communication connections 216 that allow communication with other computing devices 218. Examples of suitable communication connections 216 include, but are not limited to: RF transmitter, receiver and / or transceiver circuitry; Universal Serial Bus (USB), parallel and / or serial ports.

[0046] As used herein, the term computer-readable medium includes computer storage medium. Computer storage medium can include volatile and non-volatile, removable and non-removable media implemented using any method or technology for storing information (e.g., computer-readable instructions, data structures, or program modules). System memory 204, removable storage device 209, and non-removable storage device 210 are examples of computer storage media (i.e., memory storage). Computer storage medium can include random access memory (RAM), read-only memory (ROM), electrically erasable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital universal disc (DVD) or other optical storage, magnetic tape, magnetic tape, disk storage or other magnetic storage devices, or any other article of manufacture that can be used to store information and is accessible by computing device 200. According to one aspect, any such computer storage medium can be part of computing device 200. Computer storage medium does not include carrier waves or other transmitted data signals.

[0047] According to one aspect, a communication medium is implemented by computer-readable instructions, data structures, program modules, or other data in a modulated data signal (e.g., a carrier wave or other transmission mechanism), and includes any information transmission medium. According to one aspect, the term "modulated data signal" describes a signal having one or more sets of characteristics or altered in a manner that encodes information in the signal. By way of example and not limitation, a communication medium includes wired media such as wired networks or direct wired connections, and wireless media such as acoustic, radio frequency (RF), infrared, and other wireless media.

[0048] In an embodiment of the invention, computing device 200 is configured to execute a hydrogen refueling station planning and operation optimization method 300 that considers the resilience of a two-tier transportation-service network. The computing device 200 includes one or more processors and one or more readable storage media storing program instructions that, when configured to be executed by the one or more processors, cause the computing device to execute the hydrogen refueling station planning and operation optimization method 300 that considers the resilience of a two-tier transportation-service network according to the embodiments of the invention.

[0049] In some embodiments, the computing device 200 that performs the hydrogen refueling station planning and operation optimization method 300 considering the resilience of the traffic-service two-layer network in this embodiment of the invention may be a terminal or a server.

[0050] The following is a detailed description of the hydrogen refueling station planning and operation optimization method 300 that takes into account the resilience of the traffic-service two-layer network in the embodiments of the present invention.

[0051] Figure 3 A flowchart illustrating a hydrogen refueling station planning and operation optimization method 300 considering the resilience of a two-layer transportation-service network, according to an embodiment of the present invention, is shown. Figure 3 As shown, the hydrogen refueling station planning and operation optimization method 300, which takes into account the resilience of the traffic-service two-layer network, includes the following steps 310-340.

[0052] Step 310: With the goal of maximizing the daily profit of the hydrogen refueling station system and minimizing the overall vulnerability of the transportation-service two-layer network (hereinafter referred to as "two-layer network"), construct an upper-layer site selection and capacity planning model.

[0053] It should be noted that all hydrogen refueling stations within the two-layer network constitute a hydrogen refueling station system. The daily profit of the hydrogen refueling station system is the difference between the total daily hydrogen sales revenue of the system and the daily construction and operation costs. The daily construction and operation costs of the hydrogen refueling station system include the investment cost of construction and the daily operating cost.

[0054] In this embodiment of the invention, the overall vulnerability of a two-layer network can be determined based on the average vulnerability of nodes, the average vulnerability of edges, and the global vulnerability of the network.

[0055] Step 320: With the goal of maximizing overall user satisfaction, construct the lower-level operating model.

[0056] In this embodiment of the invention, overall user satisfaction can be determined based on time satisfaction value and economic satisfaction value. Here, time satisfaction value is used to characterize the quality of hydrogen refueling service in the two-layer network, and economic satisfaction value is used to characterize the economic cost of hydrogen refueling for users.

[0057] Step 330: Based on the upper-level site selection and capacity planning model and the lower-level operation model, a two-level optimization model for hydrogen refueling station planning and operation can be obtained.

[0058] Step 340: The biblical optimization method based on Wasserstein distance combined with the MAPPO algorithm can be used to solve the two-level optimization model for hydrogen refueling station planning and operation, so as to obtain the optimal site and capacity scheme and the optimal demand flow allocation scheme.

[0059] It should be noted that the optimal site selection and capacity allocation scheme is used to indicate whether each candidate site is selected as a hydrogen refueling station, the construction capacity of each hydrogen refueling station, and whether each hydrogen refueling station serves each demand point. The optimal demand flow allocation scheme is used to indicate the amount of hydrogen demand that each hydrogen refueling station can meet for each demand point at each time period.

[0060] In some embodiments, the specific method for constructing the upper-level site selection and capacity planning model is as follows: First, the upper-level first objective function can be constructed with the goal of maximizing the daily profit of the hydrogen refueling station system; the upper-level second objective function can be constructed with the goal of minimizing the overall vulnerability of the two-layer network; and then, the upper-level site selection and capacity planning model can be constructed based on the first objective function and the second objective function.

[0061] The decision variables in the upper-level location and capacity planning model include the location of hydrogen refueling stations, the construction capacity of hydrogen refueling stations, and the service relationship between hydrogen refueling stations and demand points. Specifically, the decision variables in the upper-level location and capacity planning model include... , , The settings are as follows.

[0062] : A 0-1 variable representing whether candidate point j is selected as a hydrogen refueling station. 0 indicates that candidate point j is not selected, and 1 indicates that candidate point j is selected as a hydrogen refueling station. The node where the existing hydrogen refueling station is located. The node where the new hydrogen refueling station is located is designated as 1. The value is 1. Specifically, as shown in formulas (10) and (11) below.

[0063] (10) (11) : A non-negative number representing the construction capacity (kg / day) of hydrogen refueling station j (the candidate site j selected as a hydrogen refueling station).

[0064] : Indicates whether hydrogen refueling station j serves demand point i.

[0065] The upper-level first objective function, constructed with maximizing the daily profit of the hydrogen refueling station system as the primary objective, is shown in the following equation: (12) In the formula, R represents the total daily hydrogen sales revenue of the hydrogen refueling station system; SC represents the daily construction investment cost of the hydrogen refueling station system; and SO represents the daily operating cost of the hydrogen refueling station system.

[0066] The total daily hydrogen sales revenue R of the hydrogen refueling station system can be determined based on the actual hydrogen demand met by each refueling station during each time period (each time period within the day) and the retail price of hydrogen during each time period, as shown in the following formula: (13) The daily construction investment cost (SC) of a hydrogen refueling station system can be determined based on the land investment cost of all hydrogen refueling stations within the two-layer network, the equipment investment cost of all hydrogen refueling stations, and the number of operating days per year.

[0067] The land investment cost for all hydrogen refueling stations equals the sum of the construction land costs for all stations. The construction land cost for each station is related to the land price and area of ​​its location. Here, the land area of ​​a hydrogen refueling station is related to its construction capacity. Hydrogen refueling stations can be divided into three levels based on the total capacity and single-tank capacity of their hydrogen storage tanks, and their construction capacity can also be divided into three levels. The land area of ​​the proposed hydrogen refueling station is determined based on the average land area of ​​each level. The specific formula is as follows: (14) (15) In the formula, For the land investment cost of all hydrogen refueling stations, The land price for the location of hydrogen refueling station j (i.e., the candidate site j selected as a hydrogen refueling station). The area occupied by hydrogen refueling station j, The construction capacity of hydrogen refueling station j; The average floor area of ​​hydrogen refueling stations at levels one, two, and three are respectively. The capacity boundaries for hydrogen refueling stations at levels one, two, and three are defined, and in some embodiments, these boundaries can be set to 8000 kg, 5000 kg, and 3000 kg, respectively.

[0068] The equipment invested in a hydrogen refueling station (on-site equipment) includes compressors, hydrogen storage tanks, and hydrogen dispensers. Therefore, the equipment investment cost for all hydrogen refueling stations includes the purchase cost of compressors, hydrogen storage tanks, and hydrogen dispensers, as shown in the following formula.

[0069] (16) (17) (18) In the formula, The cost of purchasing compressors for all hydrogen refueling stations, The cost of purchasing hydrogen storage tanks for all hydrogen refueling stations. Cost of purchasing hydrogen refueling machines for all hydrogen refueling stations; , , These are the unit prices for the compressor, hydrogen storage tank, and hydrogen refueling machine, respectively. , , These are the rated daily hydrogen processing capacities of the compressor, hydrogen storage tank, and hydrogen refueling machine, respectively.

[0070] Furthermore, the daily construction investment cost SC of the hydrogen refueling station system can be determined based on the land investment cost of all hydrogen refueling stations within the two-layer network, the equipment investment cost of all hydrogen refueling stations (the sum of the purchase costs of compressors, hydrogen storage tanks, and hydrogen dispensers for all hydrogen refueling stations), and the number of operating days per year, as shown in the following formula: (19) In the formula, This refers to the number of operating days per year for a hydrogen refueling station.

[0071] The daily operating cost SO of a hydrogen refueling station system includes the total daily cost of hydrogen procurement, the total daily cost of equipment operation and maintenance, and the total daily cost of labor.

[0072] The total daily hydrogen procurement cost can be determined based on the hydrogen demand met by each hydrogen refueling station at different times and the unit price of hydrogen procurement, as shown in the following formula: (20) In the formula, This represents the total daily hydrogen procurement cost (RMB / day). The unit price of hydrogen during the τ period is expressed in yuan / kg. For the decision variables of the lower-level operation model, it represents the amount of hydrogen demand (kg) generated by demand point i in time period t and satisfied by hydrogen refueling station j in time period τ. For upper-level decision variables, it represents the service affiliation between demand point i and hydrogen refueling station j (whether hydrogen refueling station j serves demand point i). Dynamic response time window for demand in period t This formula ensures that hydrogen procurement costs are incurred only along the planned service routes.

[0073] The total daily equipment maintenance cost is the sum of the daily equipment maintenance costs of all hydrogen refueling stations within the two-tier network, as shown in the following formula. The daily equipment maintenance cost of each hydrogen refueling station includes the daily maintenance costs (mainly maintenance fees and depreciation) of all equipment (compressors, hydrogen storage tanks, and hydrogen dispensers) within that station.

[0074] (twenty one) In the formula, Total daily equipment maintenance cost; The daily operation and maintenance cost of equipment at hydrogen refueling station j.

[0075] The total daily labor cost is the sum of the daily labor costs of all hydrogen refueling stations within the two-layer network, as shown in the following formula: (twenty two) In the formula, Total cost of Japanese workers; The daily labor cost for hydrogen refueling station j.

[0076] The daily operating cost SO of a hydrogen refueling station system is equal to the sum of the daily total cost of hydrogen procurement, the daily total cost of equipment operation and maintenance, and the daily total cost of labor, as shown in the following formula: (twenty three) In some embodiments, the ideal point method (Euclidean distance method) can be used to integrate the above three vulnerability indicators (average node vulnerability, average edge vulnerability, and global network vulnerability) into a comprehensive vulnerability indicator. That is, the ideal point method (Euclidean distance method) can be used to obtain the comprehensive vulnerability of a two-layer network based on the average node vulnerability, average edge vulnerability, and global network vulnerability of the two-layer network.

[0077] In step 210, the ideal point method (Euclidean distance method) can be used. Based on the average vulnerability of nodes, the average vulnerability of edges, and the global vulnerability of the two-layer network, the upper-layer second objective function is constructed with minimizing the overall vulnerability of the two-layer network as the upper-layer second objective. The upper-layer second objective function is shown in the following formula: (twenty four) In the formula, F2 represents the overall vulnerability of the two-layer network; This represents the average vulnerability of nodes in a two-layer network. This represents the average fragility of the edges in a two-layer network. This indicates the global vulnerability of a two-layer network.

[0078] Based on formulas (1)-(9) above, it can be understood that the average vulnerability of nodes in a two-layer network can be determined based on the vulnerability of each hydrogen refueling station. The vulnerability of each hydrogen refueling station can be obtained by normalizing its weighted clustering coefficient, which is related to its original weight. The average vulnerability of edges in a two-layer network can be determined based on the vulnerability of the paths between each demand point and each hydrogen refueling station in the two-layer network. This vulnerability is related to the probability of path interruption between each demand point and each hydrogen refueling station and the change in the average path length caused by path interruption. The global vulnerability of a two-layer network can be obtained by normalizing the global network efficiency.

[0079] In some embodiments, when constructing the upper-level site selection and capacity planning model, multiple upper-level constraints also need to be considered. That is, the upper-level site selection and capacity planning model can be constructed based on a first objective function, a second objective function, and multiple upper-level constraints. These multiple upper-level constraints include: site coverage constraints, demand satisfaction constraints, logical association constraints, vulnerability threshold constraints, decision variable domain constraints, capacity standardization constraints, capacity range constraints, land area constraints, two-layer network scale constraints, and hydrogen refueling station configuration constraints.

[0080] Specifically, the site coverage constraint means that each demand point is covered by at least one constructed hydrogen refueling station to ensure that all areas have access to hydrogen refueling services. The specific formula is as follows: (25) (26) In the formula, It is the coverage coefficient of hydrogen refueling station j to demand point i. The coverage coefficient is 1 when the distance between the demand point and the hydrogen refueling station is within the service radius of the hydrogen refueling station, and 0 otherwise. This is the maximum allowable distance from the demand point to the hydrogen refueling station j, i.e., the service radius (coverage area) of the hydrogen refueling station j. Let be the distance between demand point i and hydrogen refueling station j.

[0081] The constraints for meeting demand include daily average hydrogen demand forecasting constraints, probability distribution constraints for the initial hydrogen refueling period in the 96-period time frame, and time-series load allocation constraints.

[0082] Among them, the daily average hydrogen demand forecasting constraint is used to predict the predicted number of hydrogen fuel cell vehicles applied to each demand point in each scenario, and to calculate the daily hydrogen demand of each demand point based on the predicted number of hydrogen fuel cell vehicles applied to each demand point in each scenario and the daily average hydrogen demand per vehicle.

[0083] 7. Specifically, the demand can be predicted using the daily hydrogen demand method. Considering only the commercial application of hydrogen fuel cell vehicles, the scenarios are divided into six categories: long-haul freight transportation, urban freight delivery, public transportation, commuter passenger transport, sanitation, and postal services. Based on the promotion plan, economic development level, and the status of the hydrogen energy industry, the predicted number of hydrogen fuel cell vehicles applied to each scenario at each demand point is predicted. The specific calculation method is shown in formula (27), and the result is rounded up to conform to policy and actual demand.

[0084] (27) In the formula, For the application scenarios of hydrogen fuel cell vehicles, Represents the scenario of long-haul freight transportation. Represents urban freight delivery scenarios. Represents a public transportation scenario. Represents commuter passenger transport scenarios, Represents the sanitation scene. Represents a postal service scenario. For use within the planned area Predicted number of hydrogen fuel cell vehicles in the scenario Let i be the annual freight volume at demand point i in a long-haul freight transportation scenario. The annual concrete output of the concrete mixing plant at point i in the urban freight distribution process. Let i be the number of bus routes at demand point i in the public transportation scenario. Let i be the number of commuter routes for demand point i in a commuter passenger transport scenario. The number of hydrogen fuel cell vehicles owned for sanitation workers, representing demand point i in the sanitation sector. For the number of hydrogen fuel cell vehicles used in postal transportation at demand point i in the postal scenario, For demand point i, apply to Predicted number of hydrogen fuel cell vehicles in the scenario.

[0085] The daily hydrogen demand at each demand point is calculated based on the predicted number of hydrogen fuel cell vehicles applied to each scenario and the average daily hydrogen demand per vehicle, as shown in formula (28).

[0086] (28) In the formula, This represents the daily hydrogen demand (kg / year) at demand point i. Indicates that demand point i is applied to The average daily hydrogen demand per vehicle in the scenario (kg / (day·vehicle)).

[0087] Constraints on the Probability Distribution of the Starting Hydrogen Refueling Time in 96 Time Periods: To simulate the randomness of user refueling behavior under real traffic flow, and referencing instantaneous refueling theory, an instantaneous hydrogen refueling probability model is introduced. This transforms random driving and refueling behavior into a continuous time-period probability distribution. The time period t at which hydrogen fuel cell vehicles enter a refueling station to begin refueling in each scenario is set to follow a normal probability density function. As shown in the following formula.

[0088] (29) In the formula, express The probability density value of a hydrogen fuel cell vehicle starting hydrogen refueling in time period t under the given scenario; This represents the expected (mean) value during the initial hydrogenation period, corresponding to the central period of peak load. The standard deviation represents the dispersion of hydrogenation behavior over time; t represents the time period variable (corresponding to 96 time periods throughout the day). Various scenarios... and The parameters can be set based on the statistical analysis results of the initial refueling time of typical hydrogen fuel cell vehicles such as buses, taxis, and official vehicles, to ensure that the prediction model matches the actual driving behavior.

[0089] Based on this, normalization is performed to obtain Standardized hydrogenation probability coefficient for time period t in the scenario To ensure consistency in the total allocation: (30) In the formula, for The standardized hydrogenation probability coefficient for time period t in the scenario is summed to 1 over the 96 time periods throughout the day.

[0090] Time-series load allocation constraint: Hydrogen demand generated at demand point i during time period t The value is obtained by weighted summation of the daily hydrogen demand at demand point i under six scenarios (the daily hydrogen demand of demand point i applied to each scenario) and the standardized hydrogen refueling probability coefficient for the corresponding time period. The specific formula is as follows: (31) In the formula, This represents the hydrogen demand (kg / 15min) generated at demand point i during time period t, which serves as the core input load for the lower-level day-ahead scheduling model.

[0091] Furthermore, the probability of meeting the hydrogen demand is greater than or equal to the given confidence level. As shown in the following formula. The confidence level in this embodiment of the invention is 0.95.

[0092] (32) Logical association constraint: Hydrogen refueling station j is allowed to serve demand point i only if candidate point j is selected as a hydrogen refueling station and demand point i is within its coverage area (the distance between demand point i and hydrogen refueling station j is within the service radius of hydrogen refueling station j), as shown in the following formula.

[0093] (33) The vulnerability threshold constraints include node vulnerability threshold constraints and edge vulnerability threshold constraints, as shown in equations (34) and (35) below: (34) In the formula, This represents the maximum tolerable threshold for node vulnerability. This node vulnerability threshold constraint is used to filter out candidate nodes in the two-layer network that are too vulnerable and unsuitable for building hydrogen refueling stations. Furthermore, the construction of hydrogen refueling stations is prohibited on high-risk nodes (nodes whose vulnerability exceeds the threshold).

[0094] (35) In the formula, This represents the maximum permissible threshold for edge vulnerability. This edge vulnerability threshold constraint is used to avoid using highly vulnerable transport paths (edge ​​vulnerability exceeding the threshold), ensuring the robustness of the two-layer network.

[0095] The domain constraints for decision variables are used to specify the type and range of values ​​for each decision variable, as shown in the following formula: (36) The capacity standardization constraint is used to limit the hydrogen refueling station capacity to an integer multiple of 100 kg / d, as shown in the following formula: (37) In the formula, It is a multiple of the capacity, and is a non-negative integer.

[0096] Capacity range constraints are used to limit the minimum and maximum construction capacity of a single hydrogen refueling station to ensure economic efficiency and operational safety. The specific formula is as follows: (38) In the formula, This represents the minimum construction capacity for a single hydrogen refueling station. This represents the maximum construction capacity for a single hydrogen refueling station.

[0097] Land area constraints: used to determine the hydrogen refueling station level and corresponding land area based on the construction capacity of the hydrogen refueling station, as shown in formula (15) above; and to limit the minimum and maximum land area of ​​a single hydrogen refueling station, as shown in formula (39) below.

[0098] (39) In the formula, The minimum footprint for a single hydrogen refueling station. This refers to the maximum floor area of ​​a single hydrogen refueling station.

[0099] The two-layer network size constraint is used to limit the minimum number of hydrogen refueling stations that can be built, as shown in the following formula: (40) In the formula, The minimum number of hydrogen refueling stations required to meet the requirements.

[0100] Hydrogen refueling station configuration constraints: Each demand point has at least k alternative hydrogen refueling stations within the coverage area of ​​the hydrogen refueling station, as shown in Equation (41) below; and, in order to ensure that the capacity is still sufficient without running the lower-level operation model, verification is required, as shown in Equation (42) below.

[0101] (41) (42) Equation (42) above shows that if =0 (hydrogen refueling station j does not serve demand point i), the constraint automatically becomes invalid. If =1 (hydrogen refueling station j serves demand point i), then the total capacity of all other hydrogen refueling stations connected to demand point i must be greater than the hydrogen demand of demand point i.

[0102] In some embodiments, in step 320, a lower-level objective function can be constructed with the goal of maximizing overall user satisfaction, and a lower-level operating model can be constructed based on the lower-level objective function.

[0103] The decision variables for the lower-level operational model are the demand flow allocation for each time period. Specifically, the decision variables for the lower-level operational model are... , represents the amount of hydrogen demand (kg) generated at demand point i in time period t that is met by hydrogen refueling station j in time period τ, where τ is the actual time period during which hydrogen refueling station refuels the hydrogen fuel cell vehicle (the actual time period for providing hydrogen demand).

[0104] In some embodiments, when constructing the lower-level objective function in step 320, the TOPSIS method can be used to obtain the overall user satisfaction based on the time satisfaction value and the economic satisfaction value.

[0105] The time satisfaction value can be calculated using the following formula: (43) In the formula, This represents the satisfaction score over time. The time-dependent satisfaction decay function is usually defined as follows: This is used to penalize delayed execution of hydrogen refueling tasks.

[0106] The economic satisfaction score can be calculated using the following formula: (44) In the formula, This represents the economic satisfaction score. The cost satisfaction decay function is defined as follows: (45) In the formula, The time-of-use retail price of hydrogen during period τ; These are the highest price during the peak period and the lowest price during the trough period, set for the entire day. This represents the price sensitivity coefficient.

[0107] The specific process of obtaining overall user satisfaction based on time satisfaction and economic satisfaction using the TOPSIS method is as follows: First, calculate the distance (regret level) between the time satisfaction value, the economic satisfaction value and the positive ideal point using the following formula. ): (46) And, calculate the distance (safety level) from the time satisfaction value, economic satisfaction value to the negative ideal point using the following formula. ): (47) It should be noted that the greater the distance to the negative ideal point, the further away from the most painful state.

[0108] Furthermore, using TOPSIS's classic relative proximity formula, the satisfaction levels of the two dimensions are integrated into a dimensionless comprehensive satisfaction index. That is, the overall user satisfaction can be determined based on the distances of the time satisfaction value and the economic satisfaction value to the positive ideal point and the distance to the negative ideal point, as shown in the following formula: (48) In the formula, To assess overall user satisfaction.

[0109] In some embodiments, when constructing the lower-level operational model based on the lower-level objective function, multiple lower-level constraints also need to be considered. That is, the lower-level operational model can be constructed based on the lower-level objective function and multiple lower-level constraints. These multiple lower-level constraints include service relationship coupling constraints, demand satisfaction upper limit constraints, instantaneous service rate constraints of hydrogen refueling stations, causal constraints on hydrogen fuel cell vehicle driving time, daily capacity constraints of hydrogen refueling stations, and load balancing constraints of hydrogen refueling stations.

[0110] Specifically, the service relationship coupling constraint is used to constrain the service affiliation (hydrogen refueling station j serves demand point i) only when the service affiliation is determined at the upper layer. And hydrogen refueling stations have been built. Scheduling traffic can only be generated under certain conditions. Specifically, it is shown in the following formula: (49) In the formula, It is a sufficiently large positive number, representing the maximum possible flow limit on the path; This indicates that candidate point j has been selected as a hydrogen refueling station.

[0111] The demand meets the upper limit constraint: This condition indicates that for each demand point in each time period, the total hydrogen demand allocated to each hydrogen refueling station (the total hydrogen demand satisfied by each hydrogen refueling station) does not exceed the actual hydrogen demand generated by that demand point in that time period. The specific formula is as follows: (50) In the formula, The amount of hydrogen demand generated at point i during time period t is derived from the time-series amortization calculation results, reflecting the instantaneous refueling probability of hydrogen fuel cell vehicles in different scenarios.

[0112] Instantaneous service rate constraint of hydrogen refueling stations: The amount of hydrogen demand that a hydrogen refueling station can provide / satisfy at any given time is limited by the physical rate of its hardware equipment (compressor, refueling nozzle, etc.). Specifically, it is shown in the following formula: (51) In the formula, The maximum refueling capacity (kg / 15min) of hydrogen refueling station j within a single time period (15min).

[0113] Causal constraint on hydrogen fuel cell vehicle driving time: This constraint requires the hydrogen fuel cell vehicle to complete the physical distance traveled before refueling can begin, ensuring the physical feasibility of the time sequence allocation. It is shown in the following equation: (52) In the formula, Let be the road network distance (km) between demand point i and hydrogen refueling station j; The average driving speed of the hydrogen fuel cell vehicle is 0.25 (km / h); 0.25 is the hour conversion factor for 15 minutes.

[0114] Daily capacity constraint for hydrogen refueling stations: This constraint stipulates that the total daily hydrogen demand met by a hydrogen refueling station must not exceed the station's construction capacity.

[0115] (53) In the formula, Let be the construction capacity (kg) of hydrogen refueling station j. This variable directly affects the construction investment cost of the upper-level site selection and capacity planning model. Through this constraint, a deep correlation can be established between the capacity configuration of the upper-level site selection and capacity planning model and the resource allocation of the lower-level operation model.

[0116] Hydrogen refueling station load balancing constraints: To avoid the blind over-concentration of hydrogen energy dispatch traffic to a few refueling stations and to prevent failures caused by local hardware overload and the risk of cascading failures, load balancing constraints for hydrogen refueling stations can be explicitly introduced into the lower-level operation model. These constraints smooth out load fluctuations between refueling stations, guiding traffic to achieve a long-term balanced distribution in the spatial network and improving operational resilience. The specific formula is as follows: (54) (55) (56) In the formula, This indicates the load factor of hydrogen refueling station j. This represents the average load level of all existing hydrogen refueling stations; it is the maximum allowable threshold for the variance of the hydrogen refueling station load rate given by the system.

[0117] Based on the upper-level site selection and capacity planning model and the lower-level operation model, a two-level optimization model for hydrogen refueling station planning and operation is obtained. In order to characterize the uncertainty of demand side and improve the resilience of the planning scheme, a bibliometric optimization method based on Wasserstein distance can be used to quantify the uncertainty of hydrogen demand distribution. Combined with the MAPPO algorithm, the two-level optimization model for hydrogen refueling station planning and operation can be solved.

[0118] Specifically, the bilayer optimization method based on Wasserstein distance can be used to perform uncertainty robust transformation on the bilayer optimization model for hydrogen refueling station planning and operation (including constructing an uncertainty set and incorporating the uncertainty of hydrogen demand distribution into the lower-level operation model) to obtain the transformed bilayer optimization model, and then the MAPPO algorithm can be used to solve the bilayer optimization model.

[0119] The following section introduces a Bruker bar optimization method based on Wasserstein distance: First, we can construct a distribution based on historical experience. Centered on, with radius The fuzzy set P contains all potential true probability distributions: (57) Subsequently, optimal transport theory can be used to measure the difference between the actual distributed fuzzy set P and the historical empirical distribution. Distance between: (58) Define random variables The spatiotemporal range of values ​​and its uncertainty budget constraints : (59) The decision variable x of the upper-level site selection and sizing planning model needs to be optimized to achieve the overall system objective under the worst-case expectation of the lower-level running model: (60) (61) The constraints of the upper-level site selection and capacity planning model include site coverage constraints, capacity range constraints, and logical association constraints.

[0120] (62) The lower-level operating model, given the decision variable x of the upper-level location and sizing planning model, and stochastic implementation Under the premise of maximizing system service utility by scheduling decision variable y: (63) This subproblem restricts the capacity configuration of the upper layer through constraint terms. The data is directly transmitted to the lower-level real-time scheduling space. (64) Since it is a bi-objective lexicographical optimization, the dual transformation of the deblog bar mainly acts on the lower-level operating model (overall user satisfaction) that contains uncertain requirements.

[0121] The formula remains unchanged, transforming the overall user satisfaction under the worst-case distribution into a model based on dual variables. Deterministic expression: (65) In some embodiments, the specific process of solving the two-level optimization model for hydrogen refueling station planning and operation in step 340 using the Wasserstein distance-based bibliometric optimization method combined with the MAPPO algorithm is as follows: Obtain the model input parameters, which include parameters of multiple candidate points and multiple demand points in the two-layer network, and historical time series (96 time periods) hydrogen demand data (including the hydrogen demand of each demand point in each time period under each historical scenario). In one embodiment, the two-layer network contains a total of 206 candidate points and 118 demand points.

[0122] Subsequently, a historical hydrogen demand empirical distribution can be constructed based on historical time-series hydrogen demand data. A biblical bar optimization method based on Wasserstein distance is adopted to construct a distributed fuzzy set centered on the historical hydrogen demand empirical distribution to quantify the uncertainty of hydrogen demand distribution. The biblical bar optimization model for hydrogen refueling station planning and operation is then transformed into a biblical bar optimization model that incorporates the uncertainty of hydrogen demand distribution.

[0123] Using a lexicographical order method for hierarchical classification: the primary objective of the upper-level site selection and capacity planning model (maximizing daily profit) can be prioritized, while the secondary objective of the upper-level site selection and capacity planning model (minimizing overall vulnerability) can be prioritized.

[0124] Furthermore, a spectral clustering algorithm can be used to aggregate multiple demand points into four demand alliances, establish an agent corresponding to each demand alliance, and configure observation and action spaces for each agent to construct a multi-agent reinforcement learning environment. Here, the multi-agent reinforcement learning environment can be encapsulated based on Gym.

[0125] Subsequently, based on a multi-agent reinforcement learning environment, each agent performs joint actions of location selection and traffic allocation according to its corresponding observation space. The value of the joint actions is centrally evaluated through a global value network (centralized training), and the policy network is stably updated through a policy pruning mechanism to iteratively solve the above-mentioned bipolar two-layer optimization model. This continues until the algorithm converges, outputting the optimal location and capacity sizing scheme and the optimal demand traffic allocation scheme.

[0126] In some embodiments, the Beijing hydrogen refueling station network is selected as a case study, and the two-layer optimization model for hydrogen refueling station planning and operation constructed in this invention is applied. A scheduling cycle is set to 24 hours, and a unit scheduling period is 15 minutes. Demand data from a typical day in Beijing is selected as the model input parameters. In this embodiment, a total of 118 demand points and 206 candidate hydrogen refueling station points are set, of which 11 candidate points are existing hydrogen refueling stations. Figure 4 A schematic diagram of the distribution of demand points according to some embodiments of the present invention is shown. Figure 5A schematic diagram showing the distribution of candidate hydrogen refueling stations according to some embodiments of the present invention is provided.

[0127] Figure 6 A schematic diagram illustrating the hydrogenation probability distribution under various scenarios according to some embodiments of the present invention is shown. For example... Figure 6 As shown, six scenarios are mapped to four typical refueling behaviors based on the similarity of vehicle operation nature and driving trajectory: urban freight distribution and postal transportation are mapped to taxis (bi-peak pattern); public transportation and sanitation scenarios are mapped to buses (refueling during the last shift); commuter passenger transport is mapped to official vehicles (fixed-time pick-up and drop-off); and long-haul freight transport is mapped to private vehicles (discrete distribution). Specifically, long-haul freight transport corresponds to 16:00-22:00 and follows a normal distribution. Freight city delivery corresponds to 02:30-05:30 and 11:30-14:30, respectively, and follows a normal distribution. and The bus scenarios, corresponding to 12:30-15:30 and 21:30-00:30 the next day, respectively follow a normal distribution. and Commuter buses operate between 17:30 and 20:30, following a normal distribution. The sanitation service hours correspond to 21:30 to 00:30 the next day, and follow a normal distribution. Postal services operate from 11:30 to 14:30 and follow a normal distribution. .

[0128] In some embodiments, to verify the effectiveness of the two-level optimization model for hydrogen refueling station planning and operation, four schemes are set up for comparative analysis. The scheme settings are shown in Table 1.

[0129] Table 1 Scheme Settings

[0130] The results of the four solutions are compared in Table 2.

[0131] Table 2 Comparison of Solution Results

[0132] As shown in Table 2, Scheme 1 prioritizes economic profit, achieving the highest daily profit but exhibiting significantly higher vulnerability than other schemes; Scheme 2 prioritizes resilience, resulting in the lowest vulnerability but a substantial decrease in profit and satisfaction; Scheme 4 prioritizes resilience, but its satisfaction rate is only 72%. In contrast, Scheme 3 prioritizes profit, balancing economic efficiency and resilience, achieving lower vulnerability than Scheme 1 and the highest satisfaction rate. This demonstrates that incorporating vulnerability into the objective function effectively avoids high-risk nodes, achieving multi-objective collaborative optimization, and validating the effectiveness of the invented two-level optimization model for hydrogen refueling station planning and operation.

[0133] After solving the problem, it was finally determined that 114 new hydrogen refueling stations should be built, thus forming a two-layer network of hydrogen refueling stations including 125 hydrogen refueling stations and 118 demand points. The daily profit of the hydrogen refueling station system in this two-layer network is 407,100 yuan, the comprehensive vulnerability is 0.30, and the comprehensive user satisfaction is 0.8429. Figure 7 A schematic diagram of the location distribution of hydrogen refueling stations according to some embodiments of the present invention is shown, in which blue indicates candidate points that were not selected as hydrogen refueling stations, and red indicates candidate points that were selected as hydrogen refueling stations.

[0134] Figure 8 A schematic diagram illustrating the service relationship between hydrogen refueling stations and points of demand according to some embodiments of the present invention is shown. For example... Figure 8 As shown, most demand points are within the economic service radius, achieving a trade-off between cost and timeliness; the many-to-many connections between hydrogen refueling stations and demand points form a network structure with capacity redundancy, possessing fault tolerance capability for single-point failures, thereby ensuring the continuity of the hydrogen supply chain.

[0135] also, Figure 9 A schematic diagram illustrating the distribution of total service traffic over different time periods is shown in some embodiments of the present invention. For example... Figure 9 As shown, hydrogen demand exhibits a multi-peak heterogeneous characteristic: four peaks are formed during both the demand generation and refueling completion periods, corresponding to the morning peak (7-9 am), midday (12-2 pm), evening peak (5-7 pm), and nighttime (9-11 pm), reflecting the high coupling between hydrogen logistics and commuting and freight scheduling. The peak value during the refueling completion period is higher than that during the demand generation period, and the peak and trough are smoother, indicating that hydrogen refueling stations have queuing buffer and capacity adjustment effects. Through time delay smoothing, dynamic matching of supply and demand is achieved, demonstrating the robustness of the model in dealing with time-varying demand.

[0136] In summary, the hydrogen refueling station planning and operation optimization method 300 of the present invention, which considers the resilience of the traffic-service two-layer network, constructs a two-layer optimization model for hydrogen refueling station planning and operation based on the establishment of a two-layer network vulnerability assessment model. This model has upper-layer objectives of maximizing daily profit and minimizing overall vulnerability, and lower-layer objectives of maximizing overall user satisfaction. The method introduces the bibliometric optimization method to characterize the uncertainty of demand distribution and combines it with the MAPPO algorithm for efficient solution. This strengthens the resilience of the hydrogen refueling station network, optimizes network layout and risk resistance, and enables the two-layer network layout to effectively connect hydrogen supply and demand, achieving synergistic optimization of economy, resilience, and service quality. Furthermore, based on the two-layer network architecture, the upper-layer service network issues scheduling instructions, and the lower-layer traffic network provides feedback on user satisfaction, forming a "planning-execution-feedback" closed loop, which effectively avoids cascading risks caused by single-point failures.

[0137] By way of example, and not limitation, readable media include readable storage media and communication media. Readable storage media stores information such as computer-readable instructions, data structures, program modules, or other data. Communication media generally embodies computer-readable instructions, data structures, program modules, or other data in the form of modulated data signals such as carrier waves or other transmission mechanisms, and includes any information delivery medium. Any combination of the above is also included within the scope of readable media.

[0138] In the specification provided herein, the algorithms and displays are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used with the examples of this invention. The required structure for constructing such systems is apparent from the above description. Furthermore, this invention is not directed to any particular programming language. It should be understood that the contents of the invention described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing the best mode of implementation of the invention.

[0139] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention 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.

[0140] Similarly, it should be understood that, in order to streamline this disclosure and aid in understanding one or more of the various aspects of the invention, in the above description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof.

[0141] Those skilled in the art will understand that modules, units, or components of the device in the examples disclosed in this invention can be arranged in the device as described in this embodiment, or alternatively, can be located in one or more devices different from the device in this example. The modules in the foregoing examples can be combined into a single module or further divided into multiple sub-modules.

[0142] Unless otherwise specified, the use of ordinal numbers such as “first,” “second,” “third,” etc., to describe ordinary objects merely indicates different instances of similar objects and is not intended to imply that the objects being described must have a given order in time, space, ordering, or any other manner.

Claims

1. A method for optimizing the planning and operation of hydrogen refueling stations, taking into account the resilience of a two-layer transportation-service network, executed in a computing device, wherein the two-layer network comprises a service network and a transportation network, the service network being adapted to consist of hydrogen refueling stations for supplying hydrogen and demand points for generating hydrogen demand, the transportation network being adapted to consist of hydrogen refueling stations, demand points, hydrogen fuel cell vehicles, and roads, and all hydrogen refueling stations within the two-layer network forming a hydrogen refueling station system, the method comprising: With the goal of maximizing the daily profit of the hydrogen refueling station system and minimizing the overall vulnerability of the two-layer network, an upper-layer site selection and capacity planning model is constructed. The daily profit of the hydrogen refueling station system is the difference between its total daily hydrogen sales revenue and its daily chemical construction and operation costs, which include both construction investment costs and daily operating costs. The overall vulnerability is determined based on the average vulnerability of nodes, the average vulnerability of edges, and the global vulnerability of the two-layer network. With the goal of maximizing overall user satisfaction, a lower-level operating model is constructed. The overall user satisfaction is determined based on time satisfaction value and economic satisfaction value. The time satisfaction value is used to characterize the hydrogen refueling service quality of the two-layer network, and the economic satisfaction value is used to characterize the economic cost of hydrogen refueling for users. Based on the upper-level site selection and capacity planning model and the lower-level operation model, a two-level optimization model for hydrogen refueling station planning and operation is obtained. A bilayer optimization method based on Wasserstein distance combined with the MAPPO algorithm is used to solve the two-layer optimization model for the planning and operation of hydrogen refueling stations, resulting in the optimal site selection and capacity setting scheme and the optimal demand flow allocation scheme. The optimal site selection and capacity setting scheme indicates whether each candidate point is selected as a hydrogen refueling station, the construction capacity of each hydrogen refueling station, and whether each hydrogen refueling station serves each demand point. The optimal demand flow allocation scheme indicates the amount of hydrogen each hydrogen refueling station can meet the hydrogen demand of each demand point in each time period.

2. The method as described in claim 1, wherein, The total daily hydrogen sales revenue is determined based on the hydrogen demand met by each hydrogen refueling station at each time period and the retail price of hydrogen. The daily construction investment cost of the hydrogen refueling station system is suitable for being determined based on the land investment cost of all hydrogen refueling stations, the equipment investment cost of all hydrogen refueling stations, and the number of operating days per year. The daily operating cost of the hydrogen refueling station system includes the total daily cost of hydrogen procurement, the total daily cost of equipment operation and maintenance, and the total daily cost of labor.

3. The method as described in claim 2, wherein, The land investment cost of all hydrogen refueling stations is equal to the sum of the construction land costs of all hydrogen refueling stations. The construction land cost of each hydrogen refueling station is related to the land price and area of ​​the location of the hydrogen refueling station. The equipment investment cost of all hydrogen refueling stations includes the purchase cost of compressors, hydrogen storage tanks, and hydrogen refueling machines for all hydrogen refueling stations. The total daily hydrogen procurement cost is determined based on the hydrogen demand and unit price of hydrogen at each hydrogen refueling station during different time periods; the total daily equipment operation and maintenance cost is the sum of the daily equipment operation and maintenance costs of all hydrogen refueling stations. The total daily labor cost is the sum of the daily labor costs of all hydrogen refueling stations.

4. The method as described in any one of claims 1-3, with the upper-level objective of maximizing the daily profit of the hydrogen refueling station system and minimizing the overall vulnerability of the two-layer network, constructs an upper-level site selection and capacity planning model, comprising: With maximizing the daily profit of the hydrogen refueling station system as the primary objective, a primary objective function is constructed. Using the ideal point method, based on the average vulnerability of nodes, the average vulnerability of edges, and the global vulnerability of the network in the two-layer network, the upper-layer second objective function is constructed with the minimum comprehensive vulnerability of the two-layer network as the upper-layer second objective. Based on the first objective function and the second objective function, an upper-level location and capacity planning model is constructed.

5. The method of claim 4, wherein, Based on the first objective function and the second objective function, an upper-level location and capacity planning model is constructed, including: Based on the first objective function, the second objective function, and various upper-level constraints, an upper-level location and capacity planning model is constructed. The various upper-level constraints include: site coverage constraints, demand satisfaction constraints, logical association constraints, vulnerability threshold constraints, decision variable domain constraints, capacity standardization constraints, capacity range constraints, land area constraints, two-layer network scale constraints, and hydrogen refueling station configuration constraints.

6. The method as described in any one of claims 1-5, The average vulnerability of the nodes is suitable for determination based on the vulnerability of each hydrogen refueling station, wherein, The vulnerability of the hydrogen refueling station is suitable for obtaining by normalizing the weighted clustering coefficient of the hydrogen refueling station; The average vulnerability of the connection is suitable for determination based on the vulnerability of the path between each demand point and each hydrogen refueling station, and is related to the probability of path interruption between each demand point and each hydrogen refueling station and the change in the average path length of the network caused by path interruption. The global vulnerability of the network is suitable for obtaining by normalizing the global network efficiency.

7. The method as described in any one of claims 1-6, wherein a lower-level operating model is constructed with the goal of maximizing overall user satisfaction, comprising: A lower-level objective function is constructed with the goal of maximizing overall user satisfaction as the lower-level objective, and a lower-level operating model is constructed based on the lower-level objective function; The overall user satisfaction is suitable for obtaining using the TOPSIS method based on time satisfaction value and economic satisfaction value.

8. The method of claim 7, wherein, Based on the aforementioned lower-level objective function, a lower-level operational model is constructed, including: Based on the aforementioned lower-level objective function and various lower-level constraints, a lower-level operational model is constructed. The various lower-level constraints include: service relationship coupling constraints, demand satisfaction upper limit constraints, instantaneous service rate constraints of hydrogen refueling stations, causal constraints of hydrogen fuel cell vehicle driving time, daily capacity constraints of hydrogen refueling stations, and load balancing constraints of hydrogen refueling stations.

9. The method according to any one of claims 1-8, wherein, A bifurcation optimization method based on Wasserstein distance combined with the MAPPO algorithm is used to solve the two-level optimization model for the planning and operation of the hydrogen refueling station, obtaining the optimal site selection and capacity allocation scheme and the optimal demand flow allocation scheme, including: Obtain model input parameters, which include parameters of multiple candidate points and multiple demand points of the two-layer network, and historical time series hydrogen demand data; Based on the historical time series hydrogen demand data, a historical hydrogen demand empirical distribution is constructed. A bibru bar optimization method based on Wasserstein distance is adopted to construct a distribution fuzzy set with the historical hydrogen demand empirical distribution as the center to quantify the uncertainty of hydrogen demand distribution. The bibru bar bibru bar optimization model for hydrogen refueling station planning and operation is transformed into a bibru bar bibru bar optimization model that embeds the uncertainty of hydrogen demand distribution. Maximizing daily profit and minimizing overall vulnerability of the aforementioned upper-level site selection and gradation planning model are taken as the primary and secondary objectives, respectively. The spectral clustering algorithm is used to aggregate the multiple demand points into four demand alliances, establish an agent corresponding to each demand alliance, and configure observation space and action space for the agent to construct a multi-agent reinforcement learning environment. Based on the multi-agent reinforcement learning environment, each agent performs joint actions of location selection and traffic allocation based on its corresponding observation space. The value of the joint actions is evaluated centrally through a global value network, and the policy network is updated through a policy pruning mechanism. This allows for iterative solution of the bilayer optimization model to obtain the optimal location and capacity setting scheme and the optimal demand traffic allocation scheme.

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

Images