Hydrogen energy full-link scheduling system and method based on big data

Abstract

The invention relates to the technical field of hydrogen energy scheduling, in particular to a hydrogen energy full-link scheduling system and method based on big data. The system comprises a hydrogen energy link data acquisition module, a hydrogen energy supply and demand prediction module, a hydrogen production storage and transportation matching module, a hydrogen energy link scheduling optimization module and a safety feedback module. Multi-dimensional operation data covering hydrogen production to terminal utilization is collected through a distributed sensor network, a supply and demand trend is predicted by adopting a stochastic differential equation and a neural random process, and a reinforcement learning optimization model is constructed in combination with link constraints to generate an initial scheduling scheme; and constructing a graph structure path planning model by using an artificial potential field algorithm to optimize full-link scheduling, and finally realizing safety evaluation and closed-loop feedback regulation and control through fuzzy logic control. According to the invention, the digital, intelligent and efficient development of the operation of each link of the hydrogen energy industry chain is promoted, and the full-link operation efficiency and intelligent level of hydrogen energy are comprehensively improved.

Description

Technical Field

[0001] The present invention relates to the field of hydrogen energy scheduling technology, and in particular to a hydrogen energy full-link scheduling system and method based on big data. Background Art

[0002] As a clean and efficient secondary energy source, hydrogen plays an increasingly important role in the global energy transition. Currently, hydrogen energy has been demonstrated in transportation, industrial energy, distributed power generation, and other fields. The industrial chain covers multiple links, including hydrogen production, storage, transportation, refueling, and end-use. As hydrogen energy applications continue to expand, the volume of data and operational complexity of each link in the hydrogen energy chain have increased significantly, placing higher demands on efficient and coordinated scheduling across the entire chain. However, the following problems still exist: the existing system has data disconnection problems between hydrogen production, storage, transportation, refueling and terminal utilization, resulting in information islands between hydrogen production, storage, transportation and terminal hydrogen use; traditional hydrogen supply and demand forecasts mostly use traditional statistical methods or simple time series analysis, which is difficult to fully capture the nonlinear coupling relationship between multi-source heterogeneous data that affects the circulation of hydrogen energy, resulting in insufficient accuracy and robustness of the prediction results; the scheduling strategies in the existing systems usually adopt fixed paths or shortest path principles, and fail to introduce graph structure modeling and dynamic avoidance strategies, making it difficult to effectively deal with the impact of bottleneck nodes, risky sections and emergencies in the link, which easily leads to local resource waste and overall low operating efficiency. Summary of the Invention

[0003] To solve the above problems, the present invention provides a hydrogen energy full-link scheduling system and method based on big data, which solves the challenges of multi-link collaboration, data heterogeneous coupling and link dynamic risks in the context of the continuous expansion of hydrogen energy application. It is about how to achieve efficient data integration and collaborative scheduling among various links in the hydrogen energy full link, improve the accuracy and robustness of supply and demand forecasts, and introduce intelligent path planning to dynamically avoid bottlenecks and risk nodes, thereby building an intelligent, efficient and safe hydrogen energy scheduling system.

[0004] To achieve the above object, the technical solution adopted by the present invention is: On the one hand, a hydrogen energy full-link scheduling system based on big data includes a hydrogen energy link data acquisition module, a hydrogen energy supply and demand prediction module, a hydrogen production, storage and transportation matching module, a hydrogen energy link scheduling optimization module and a safety feedback module, which are sequentially connected in communication; The hydrogen energy link data acquisition module is used to collect multi-dimensional operating data of the entire hydrogen energy link in real time through a distributed sensor network; the entire link includes hydrogen production, compression, storage, transportation, filling and terminal utilization; The hydrogen energy supply and demand forecasting module is used to dynamically evolve and predict the spatiotemporal dynamic change trends of the hydrogen supply and demand sides based on the multi-dimensional operating data using a hydrogen energy supply and demand forecasting model constructed by stochastic differential equations and neural stochastic processes, and output supply and demand forecast results; The hydrogen production, storage and transportation matching module is used to extract link operation constraints based on the supply and demand forecast results and the multi-dimensional operation data, and generate an initial scheduling plan using a multi-objective joint optimization strategy driven by reinforcement learning; The hydrogen energy link scheduling optimization module is used to construct a hydrogen energy link scheduling model based on the initial scheduling plan using a graph structure path planning method based on an artificial potential field algorithm, dynamically avoid bottleneck nodes, congested sections, and high-risk areas in the transportation path, and generate an optimal scheduling instruction set for the entire link in combination with scheduling timing and resource cost constraints; The safety feedback module is used to collect real-time monitoring data based on the full-link optimal scheduling instruction set, use fuzzy logic control rules to perform safety assessment and intervention decisions on link operation, identify abnormal conditions and trigger closed-loop control feedback, and output safety intervention instructions, including scheduling correction suggestions, risk warning information and safety control parameter adjustment instructions.

[0005] Furthermore, the multi-dimensional operating data includes hydrogen production, purity, pressure, temperature, remaining capacity of the storage tank, transportation route status, filling flow rate, terminal load demand, environmental parameters and equipment operating status.

[0006] Furthermore, the operation process of the hydrogen energy supply and demand forecasting module includes the following steps: Performing preprocessing operations on the multidimensional operating data, including aggregation, performing anomaly removal, missing completion, and spatiotemporal feature normalization, and reconstructing the multidimensional operating data into a structured, time-series continuous input feature matrix; Based on the input feature matrix, a hydrogen supply and demand forecasting model is constructed by using stochastic differential equations and neural stochastic processes to characterize the dynamic evolution characteristics of hydrogen supply and demand; The input feature matrix is ​​input into the hydrogen energy supply and demand forecasting model, and numerical integration and neural network-driven random process deduction are used to dynamically simulate and predict the spatiotemporal dynamic change trends of the hydrogen supply and demand sides in each region, and output supply and demand forecast results, including hydrogen supply and demand curves, corresponding confidence intervals, supply and demand trends in each region in each future time period, supply and demand gaps, predicted confidence intervals and potential risk sections.

[0007] Furthermore, the process of constructing the hydrogen energy supply and demand forecasting model includes the following steps: Based on the multi-dimensional operating data, a stochastic differential equation containing a state variable coupling term and a noise driving term is constructed to simulate the random disturbance characteristics and dynamic coupling behavior in the hydrogen supply and demand process; A neural stochastic process framework is introduced, and the solution space of the stochastic differential equation is used as the prior structure. Through joint training on historical supply and demand data and related environmental characteristics, the nonlinear time series distribution law of the supply and demand curves of each region is learned; Dynamically adjusting the kernel function parameters and prior distribution in the neural stochastic process using the maximum a posteriori estimation method, structurally coupling the stochastic differential equations with the neural stochastic process model to form a hydrogen energy supply and demand prediction model; Based on the hydrogen energy supply and demand forecasting model, the multi-dimensional operation data is input, a numerical solver is used to deduce the future trajectory of supply and demand distribution, and a supply and demand forecasting result is output.

[0008] Furthermore, the formula of the hydrogen energy supply and demand forecasting model is as follows: in, Represents the hydrogen supply and demand state function of region a The first-order derivative relative to time T, that is, the direction and intensity of changes in the hydrogen supply and demand balance trend in the region; represents the hydrogen supply and demand state function of region a at time T; T represents the time step; Indicates that region b at the historical moment Multi-dimensional operation data; Represents a nonlinear feature mapping function based on a neural network; The kernel function representing the time sensitivity of the weighted integral; represents the nonlinear coupling response function of region b to region a; The static weight coefficient representing the inter-regional coupling strength; represents the system disturbance term of region a at time t; represents the second-order derivative of the supply and demand status of region a over time; N represents the total number of regions or hydrogen energy link nodes participating in the joint modeling; Represents the hydrogen supply and demand state function of region a The second derivative with respect to time T, i.e. the acceleration of the change in supply and demand conditions; Indicates the dynamic response inertia coefficient of area a.

[0009] Furthermore, the operation process of the hydrogen production, storage and transportation matching module includes the following steps: Based on the multi-dimensional operation data and supply and demand forecast results, link operation constraints are extracted, including the production capacity boundary of the hydrogen production device, the capacity limit of the hydrogen storage container, the accessibility map of the transportation path, the service time window and priority level of the refueling node; The link operation constraints are combined to construct a joint optimization goal, which includes minimizing hydrogen production energy consumption, minimizing storage and transportation delays, maximizing task completion rate, and avoiding path congestion; Based on the joint optimization objective, a reinforcement learning framework is constructed that integrates a hierarchical state encoder and a multi-strategy behavior network. The dynamic supply and demand forecast results are used as state inputs, and the link resource scheduling cost and the supply and demand matching deviation are used as joint reward signals. The policy network is trained using a policy gradient optimization method to output a candidate set of scheduling solutions for the optimal solution domain. The Pareto optimal solution under the current constraints is screened from the candidate set of scheduling schemes to generate a structured initial scheduling scheme; the initial scheduling scheme includes the start and stop timing of hydrogen production nodes, hydrogen storage container allocation, transportation route configuration and priority sorting of refueling tasks.

[0010] Furthermore, the operation process of the hydrogen energy link scheduling optimization module includes the following steps: Based on the initial scheduling plan and in combination with the multi-dimensional operation data, a graph structure path planning method based on an artificial potential field algorithm is used to construct a hydrogen energy link scheduling model. The hydrogen energy link scheduling model uses hydrogen energy link nodes as vertices and transportation and transshipment paths as edges. Node attributes include resource capacity, task queue, and safety status. Edge attributes include path accessibility, latency, transportation risk level, and historical congestion records. Based on the hydrogen energy link scheduling model, attractive and repulsive potentials are applied to each transport route to identify and mark potential bottleneck sections and high-risk nodes in the route in real time; the attractive potential is used to guide the scheduled object to approach the target node, and the repulsive potential is used to dynamically avoid bottleneck nodes, congested sections and high-risk areas; Using a graph-structured path search method, under the influence of an artificial potential field, the transport path of each scheduling task is dynamically and iteratively adjusted, scheduling timing is optimized, bottleneck sections and high-risk nodes are proactively avoided, and the optimal scheduling plan for the entire link is generated based on the actual link resource status and scheduling task requirements. The optimal scheduling scheme is encoded to generate a full-link optimal scheduling instruction set, which includes a task execution schedule, an allocation map of transportation and storage resources, a path adjustment and switching strategy, and a key node monitoring priority sorting.

[0011] Furthermore, the formula of the hydrogen energy link scheduling model is as follows: in, It represents the value of the system's scheduling guidance function from node i to node j at time t; represents the spatial gradient operator of the artificial potential field in the path network; represents the attractive potential function; represents the repulsive potential function; represents the disturbance term on the path; represents the current task load intensity of node j at time t; represents the maximum resource service capability of node j; represents the path scheduling coupling strength from source node i to target node j through relay node k; represents the current remaining available resource ratio of relay node k; M represents the total number of relay nodes that can participate in cooperative scheduling in the current path; represents the risk level of the path from node i to node j; Indicates the historical congestion level from node i to node j; and represents the weight coefficient of the attractive potential term; Represents the weight coefficient of the repulsive potential term.

[0012] On the other hand, a hydrogen energy full-link scheduling method based on big data includes the following steps: Through the distributed sensor network, multi-dimensional operation data of the entire hydrogen energy chain is collected in real time; Based on the multi-dimensional operating data, a hydrogen energy supply and demand forecasting model constructed by stochastic differential equations and neural stochastic processes is used to dynamically evolve and forecast the spatiotemporal dynamic change trends of hydrogen supply and demand, and output supply and demand forecast results; Based on the supply and demand forecast results and in combination with the multi-dimensional operation data, link operation constraints are extracted, and a multi-objective joint optimization strategy driven by reinforcement learning is adopted to generate an initial scheduling plan; Based on the initial scheduling plan, a graph-structured path planning method based on an artificial potential field algorithm is used to construct a hydrogen energy link scheduling model. Dynamic avoidance calculations are performed on bottleneck nodes, congested sections, and high-risk areas in the transportation path. Combined with scheduling timing and resource cost constraints, an optimal scheduling instruction set for the entire link is generated. Based on the full-link optimal scheduling instruction set, real-time monitoring data is collected, and fuzzy logic control rules are used to conduct safety assessment and intervention decisions on link operation, identify abnormal conditions and trigger closed-loop control feedback, and output safety intervention instructions.

[0013] The beneficial effects of the present invention are: The present invention uses a hydrogen energy link data acquisition module, and the system realizes real-time collection of multi-dimensional operation data of the entire link, such as hydrogen production, compression, storage, transportation, filling and terminal utilization, based on a distributed sensor network. This significantly improves the system's spatiotemporal perception granularity and response speed to the link status, and provides a reliable data foundation for subsequent modules. The hydrogen energy supply and demand prediction module introduces stochastic differential equations and neural random processes to construct a joint modeling mechanism, which can characterize the complex nonlinear evolution characteristics of both ends of hydrogen supply and demand, effectively improve the prediction accuracy of hydrogen energy supply and demand trends in different regions and different time scales, and support forward-looking resource allocation decisions. The hydrogen production, storage and transportation matching module integrates supply and demand prediction results with multi-dimensional operation data, extracts link operation constraints, and adopts a multi-objective joint optimization strategy driven by reinforcement learning. It can dynamically balance between multiple objectives (such as energy efficiency, economy, and safety), effectively improving the efficiency of hydrogen energy resource allocation and the adaptability of scheduling strategies. The hydrogen energy link scheduling optimization module utilizes an artificial potential field algorithm and graph-structured path planning methods to dynamically identify and avoid bottleneck nodes, congested sections, and high-risk areas along transportation routes. It generates optimal scheduling instructions based on scheduling timelines and resource cost constraints, significantly improving the intelligence and robustness of scheduling strategies. The safety feedback module collects operational data in real time and combines it with fuzzy logic control rules for safety assessment and decision-making intervention. It can identify abnormal conditions in real time and generate safety intervention instructions, including scheduling corrections, risk warnings, and parameter adjustments. This creates a closed-loop control system encompassing prediction, response, and adjustment, significantly enhancing the safety and stability of system operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a module schematic diagram of a big data-based hydrogen energy full-link scheduling system of the present invention.

[0015] Figure 2 It is a flowchart of the operation process of the hydrogen energy link scheduling optimization module provided by one embodiment of the present invention.

[0016] Figure 3 It is a flow chart of a hydrogen energy full-link scheduling method based on big data in the present invention. DETAILED DESCRIPTION

[0017] See also Figure 1-3 As shown, the present invention relates to a hydrogen energy full-link scheduling system and method based on big data.

[0018] Example 1 A hydrogen energy full-link scheduling system based on big data, including a hydrogen energy link data acquisition module, a hydrogen energy supply and demand prediction module, a hydrogen production, storage and transportation matching module, a hydrogen energy link scheduling optimization module and a safety feedback module, which are sequentially connected in communication; The hydrogen energy link data acquisition module is used to collect multi-dimensional operating data of the entire hydrogen energy link in real time through a distributed sensor network; the entire link includes hydrogen production, compression, storage, transportation, refueling and terminal utilization; the multi-dimensional operating data includes hydrogen production, purity, pressure, temperature, remaining capacity of the storage tank, transportation route status, refueling flow rate, terminal load demand, environmental parameters and equipment operating status.

[0019] Specifically, in the hydrogen production, compression, storage, transportation, filling, and terminal hydrogen use links, multi-source sensing terminals are deployed according to process requirements and operation characteristics as follows: Hydrogen yield / purity analyzer: placed in the hydrogen production and compression unit to achieve high-precision online monitoring of gas flow and purity (through mass spectrometry or chromatography analysis).

[0020] Pressure / temperature transmitters: Deployed along key points such as pipelines, storage tanks, transport vehicles, and filling machines to support high-frequency data collection.

[0021] Tank level / capacity sensor: uses float, ultrasonic or capacitive principles to monitor changes in liquid or high-pressure gaseous hydrogen reserves in real time.

[0022] Flow meter: used at key flow nodes to monitor the instantaneous and cumulative flow rates of filling stations, transport pipelines / vehicles, and terminal hydrogen supply pipelines.

[0023] Transport route and positioning module: All transport vehicles / mobile equipment are equipped with Beidou / GPS dual-mode positioning and 4G / 5G communication modules to upload location and route status in real time.

[0024] Environmental monitors: such as temperature and humidity, wind speed, and air quality (AQI) sensors, are deployed in key environmental areas for hydrogen storage, transportation, refueling, and use.

[0025] Smart cameras / machine vision terminals: used for video collection and anomaly identification in key links, such as leakage in liquid hydrogen storage areas, abnormal open flames, and people entering by mistake.

[0026] The data types and acquisition parameters are as follows: Hydrogen production: Real-time flow rate of electrolyzer (Nm 3 / h), 1Hz refresh; Hydrogen purity: online gas chromatography (>99.999%), 5 minutes / time; Pressure / temperature: high pressure gauge (0-100MPa), thermocouple (-40~120℃), 1Hz refresh rate; Remaining capacity of storage tank: based on liquid level meter / mass flow meter, 10 seconds / time; Transport status: location coordinates (once every 5 seconds), remaining hydrogen volume, driving speed, route deviation alarm; Filling flow rate: filling gun flow meter (L / min), single filling cumulative amount; Terminal load: The terminal uses a hydrogen terminal smart meter to collect actual hydrogen usage and short-term load forecasts every 10 seconds; Environmental parameters: temperature, humidity, wind speed, gas concentration, etc., updated every 30 seconds; Equipment operating status: The equipment controller collects start / stop, fault code, current and voltage, vibration acceleration, etc., once per second.

[0027] All data is aggregated through the local edge gateway, cleaned, denoised, and initially detected for anomalies, and then reported to the main data center in real time via the 5G / wired network, providing a highly reliable, low-latency data foundation for subsequent model training and scheduling decisions.

[0028] The hydrogen energy supply and demand forecasting module is used to dynamically evolve and predict the spatiotemporal dynamic change trends of the hydrogen supply and demand sides based on the multi-dimensional operating data using a hydrogen energy supply and demand forecasting model constructed by stochastic differential equations and neural stochastic processes, and output supply and demand forecast results; The operation process of the hydrogen energy supply and demand forecasting module includes the following steps: Performing preprocessing operations on the multidimensional operating data, including aggregation, performing anomaly removal, missing completion, and spatiotemporal feature normalization, and reconstructing the multidimensional operating data into a structured, time-series continuous input feature matrix; Based on the input feature matrix, a hydrogen supply and demand forecasting model is constructed by using stochastic differential equations and neural stochastic processes to characterize the dynamic evolution characteristics of hydrogen supply and demand; The input feature matrix is ​​input into the hydrogen energy supply and demand forecasting model, and numerical integration and neural network-driven random process deduction are used to dynamically simulate and predict the spatiotemporal dynamic change trends of the hydrogen supply and demand sides in each region, and output supply and demand forecast results, including hydrogen supply and demand curves, corresponding confidence intervals, supply and demand trends in each region in each future time period, supply and demand gaps, predicted confidence intervals and potential risk sections.

[0029] The process of constructing the hydrogen energy supply and demand forecasting model includes the following steps: Based on the multi-dimensional operating data, a stochastic differential equation containing a state variable coupling term and a noise driving term is constructed to simulate the random disturbance characteristics and dynamic coupling behavior in the hydrogen supply and demand process; Specifically, the key state variables of the hydrogen supply side (such as hydrogen production rate, tank balance, and hydrogen station filling capacity) and the state variables of the demand side (such as terminal load, electric vehicle hydrogen refueling frequency, and backup hydrogen storage requests) are set; coupling factors reflecting "supply and demand interactive feedback" are introduced into the differential equation structure, such as hydrogen production response delays and hydrogen storage facility peak-shaving caused by changes in terminal demand; random disturbance sources are set, such as external temperature changes, electricity price fluctuations, traffic congestion and other factors that affect the disturbance response of the supply and demand system, and are modeled as Brownian motion driven terms; different regions construct their own differential models, and the parameters are initialized according to the regional wind speed, load density, and infrastructure capacity to support personalized dynamic simulation.

[0030] A neural stochastic process framework is introduced, and the solution space of the stochastic differential equation is used as the prior structure. Through joint training on historical supply and demand data and related environmental characteristics, the nonlinear time series distribution law of the supply and demand curves of each region is learned; Specifically, the Latent Neural Process (LNP) is used as the modeling framework, which has the ability to process uncertain inputs and output distribution predictions; the inputs include the ambient temperature and humidity at the current time point, load requests, road traffic flow, historical supply and demand fluctuations, etc.; the output is a family of hydrogen supply and demand function curves in the current area, including confidence bands, rates of change, and periodic characteristics; by performing multiple sampling training on normalized historical data, the model can generalize to input perturbations in different scenarios; dual-channel convolution (time channel + regional channel) is used to jointly encode the input data to improve the modeling ability of long-term dependencies and cross-regional relationships.

[0031] Dynamically adjusting the kernel function parameters and prior distribution in the neural stochastic process using the maximum a posteriori estimation method, structurally coupling the stochastic differential equations with the neural stochastic process model to form a hydrogen energy supply and demand prediction model; Specifically, the system regularly analyzes the latest operational and supply-demand data, dynamically and adaptively adjusting key parameters such as the kernel function parameters of the neural stochastic process model and the noise intensity of the stochastic differential equation. Each region independently optimizes parameters based on historical errors and recent trends to ensure forecast accuracy. When the external environment or link status changes significantly (such as a surge in demand or equipment fluctuations), the model automatically triggers a rapid tuning process, adjusting the neural network weights and prior distribution to maintain model adaptability. Combining historical data with the latest observations, the system automatically employs methods such as maximum a posteriori estimation to optimize model parameters in real time, recording all adjustments for easy traceability.

[0032] An efficient interface is established between the stochastic differential equation submodel and the neural stochastic process model, allowing state variables (such as production output, demand fluctuations, storage pressure, etc.) and their noise characteristics to be synchronized in real time between the submodels. The neural stochastic process framework uses the dynamic state of the stochastic differential equation solution space as a priori input to achieve model co-evolution.

[0033] During each periodic model update, the system first uses historical data to preliminarily calibrate the parameters of the stochastic differential equation and the neural process kernel function, respectively. This process then enters the joint training phase. During this phase, the model automatically calculates the overall prediction error and component contributions, and optimizes the weight distribution of the two models based on the gradient of the global loss function, achieving a deep fusion of the advantages of "physical constraints + data-driven."

[0034] If extreme abnormal events occur during actual operation (such as equipment failures or extreme market fluctuations), the system can automatically identify and temporarily adjust the model structure (such as adding high-order disturbance terms or activating emergency response networks). The structural coupling module has self-diagnostic capabilities and automatically analyzes the synergy between sub-models. If the predictive ability of a sub-model is found to have declined, the dominant role of another model is temporarily enhanced to ensure the stability and robustness of the overall prediction.

[0035] The entire process of dynamic parameter adjustment and structural coupling is automatically executed by the backend engine, requiring no human intervention. The system automatically generates visual reports of all parameter adjustments, model structure evolution, and collaborative prediction results, providing scheduling and maintenance personnel with timely tracing and decision-making reference. The system automatically records key parameter adjustments and significant model structure changes, meeting compliance and security management requirements.

[0036] Based on the hydrogen energy supply and demand forecasting model, the multi-dimensional operation data is input, a numerical solver is used to deduce the future trajectory of supply and demand distribution, and a supply and demand forecasting result is output.

[0037] Specifically, the latest full-link multi-dimensional operation data from the day is used as the prediction benchmark; a numerical integrator (such as the improved Euler-Maruyama method) is used to perform multi-step iterative deduction of the differential equation; Confidence interval output: Output upper and lower confidence limits for each predicted time point; if the confidence interval is too wide or there is significant fluctuation, the system identifies it as a potential unstable point; Supply-demand gap calculation: Real-time identification of gap sections where hydrogen supply is less than demand, and marking of the gap size; the gap sections are transmitted as input to the subsequent "hydrogen production, storage and transportation matching module" for peak shaving; Identification of high-risk areas: If the forecast shows that there will be a continuous gap in a certain period of time in the future and it cannot be quickly compensated through storage and transportation means, the system will automatically mark it as a high-risk section; the high-risk area will be linked to the "Safety Feedback Module" to indicate possible supply and demand imbalances or abnormal scheduling pressures.

[0038] Furthermore, the formula of the hydrogen energy supply and demand forecasting model is as follows: in, Represents the hydrogen supply and demand state function of region a The first-order derivative relative to time T, that is, the direction and intensity of changes in the hydrogen supply and demand balance trend in the region; when the value is positive, it indicates that supply is increasing or demand is decreasing, that is, the hydrogen surplus is expanding; when the value is negative, it indicates that supply is decreasing or demand is increasing, that is, hydrogen tension is intensifying; when the value is close to zero, the system is in a relatively balanced state, suitable for scheduling lock or energy-saving operation; represents the hydrogen supply and demand state function of region a at time T, which is used to measure the difference between the total supply and total demand of hydrogen in the region per unit time, with the unit being kilograms per hour (kg / h); T represents the time step; Indicates that region b at the historical moment Multi-dimensional operation data; Represents a nonlinear feature mapping function based on a neural network; The kernel function representing the time sensitivity of the weighted integral; It represents the nonlinear coupling response function of region b to region a, reflecting the mutual influence of supply and demand status among multiple regions. It is particularly suitable for handling scenarios such as distributed hydrogen refueling stations, cross-regional transportation, and asynchronous scheduling. The static weight coefficient representing the inter-regional coupling strength; represents the system disturbance term of region a at time t, which is used to model unpredictable fluctuation factors, such as sudden equipment failure, external environmental interference, or sampling anomalies; The second-order derivative term represents the time-varying supply and demand state of region a. This term is used to simulate the inertial characteristics of the physical processes within the link, such as the response delay of hydrogen storage tank deflation, the dynamic hysteresis caused by long-distance transportation, or the system adjustment inertia after a sudden change in terminal load. N represents the total number of regions or the number of hydrogen energy link nodes participating in the joint modeling. Represents the hydrogen supply and demand state function of region a The second derivative with respect to time T, i.e. the acceleration of the change in supply and demand conditions; It represents the dynamic response inertia coefficient of area a, which is used to adjust the system's sensitivity to the acceleration of supply and demand changes.

[0039] The calculation formula is as follows: in, Represents the time sensitivity kernel function of the weighted integral, that is, the time sensitivity of region b to region a at the historical moment The time decay weight of It represents the time sensitivity coefficient, which is used to control the window width of information retention and can be estimated based on the regional transportation distance and hydrogen storage response delay.

[0040] The calculation formula is as follows: in, The static weight coefficient representing the inter-regional coupling strength, i.e., the static impact strength of region b on the supply and demand status changes in region a; It represents the average hydrogen flow rate transported from area b to area a per historical unit time; It represents the average hydrogen flow rate transported from area c to area a per historical unit time; Represents the hydrogen quality consistency factor, which is calculated based on hydrogen purity and temperature.

[0041] The hydrogen production, storage and transportation matching module is used to extract link operation constraints based on the supply and demand forecast results and the multi-dimensional operation data, and generate an initial scheduling plan using a multi-objective joint optimization strategy driven by reinforcement learning; The operation process of the hydrogen production, storage and transportation matching module includes the following steps: Based on the multi-dimensional operation data and supply and demand forecast results, link operation constraints are extracted, including the production capacity boundary of the hydrogen production device, the capacity limit of the hydrogen storage container, the accessibility map of the transportation path, the service time window and priority level of the refueling node; Specifically, the system automatically reads the latest multi-dimensional operating data (such as production, inventory, equipment status, and transportation resources) and supply and demand forecast results (regional demand and supply in each time period).

[0042] Extract key operational constraints for each link, including: Hydrogen production capacity boundaries: the upper and lower limits of the production capacity, planned maintenance time, and real-time availability of each hydrogen production unit.

[0043] Hydrogen storage capacity limit: the maximum storage capacity, minimum safe capacity, current inventory and safety threshold of each hydrogen storage container.

[0044] Transport route accessibility: Collect information on road conditions, pipeline accessibility, real-time traffic, and the status of transport resources (vehicles / tankers / pipeline sections) to create an accessibility map.

[0045] Refueling node constraints: service time window of hydrogen refueling station (such as daily operating time period), single service capacity, priority of each node (such as emergency, key users, etc.).

[0046] The link operation constraints are combined to construct a joint optimization goal, which includes minimizing hydrogen production energy consumption, minimizing storage and transportation delays, maximizing task completion rate, and avoiding path congestion; Specifically, the system organizes the above constraints, combines them with actual business needs, and defines multiple optimization goals: Minimizing hydrogen production energy consumption: Based on energy consumption models and peak-valley electricity prices, priority is given to arranging low-energy consumption production periods and equipment.

[0047] Minimize storage and transportation delays: Prioritize short-distance, high-speed, and low-congestion storage and transportation routes and nodes to optimize the timeliness of the overall scheduling link.

[0048] Maximize task completion rate: Ensure that all tasks covered by supply and demand forecasts can be promptly allocated to corresponding link resources, improving matching and fulfillment rates.

[0049] Path congestion avoidance: Dynamically monitor the real-time congestion and risk levels of paths, and prioritize avoiding high-risk, easily congested, and maintenance-prone routes.

[0050] The above optimization goals form a quantifiable indicator system to support subsequent intelligent algorithm optimization.

[0051] Based on the joint optimization objective, a reinforcement learning framework is constructed that integrates a hierarchical state encoder and a multi-strategy behavior network. The dynamic supply and demand forecast results are used as state inputs, and the link resource scheduling cost and the supply and demand matching deviation are used as joint reward signals. The policy network is trained using a policy gradient optimization method to output a candidate set of scheduling solutions for the optimal solution domain. Specifically, the system automatically summarizes the detailed status of each link in the current link, including the start and shutdown status of hydrogen production equipment, the remaining capacity of the storage tank, the current location and idleness of the transport vehicle / pipeline, the queue length of the filling station, the task waiting list and the predicted supply and demand data for each time period.

[0052] The hierarchical state encoding is as follows: The first layer focuses on the real-time operating status of a single node (such as a hydrogen production device, a storage tank, or a transport vehicle), encoding its production capacity, availability, load, maintenance status, etc. The second layer focuses on the link level, encoding resource distribution, task queues, congestion and risk distribution, service time windows and other overall characteristics; After the state information is fused, it is input into the policy network through the feature vector to ensure that the scheduling decision takes into account both local and global perspectives.

[0053] Independent strategy branches are designed for different decision dimensions (hydrogen production, storage, transportation, and refueling), outputting specific operational recommendations (such as adjusting production, switching storage tanks, allocating transportation routes, and sequencing refueling tasks). The network structure supports the parallel deduction of multiple feasible paths, enhancing adaptability to complex scenarios. The input is an encoded high-dimensional state vector, and the output is a set of executable resource scheduling actions and their probability distributions for the system to sample and select.

[0054] A multi-objective reward system is established: reduced energy consumption and costs bring positive rewards, while overproduction, overtime, and high energy consumption result in penalties. Unfinished tasks, unmet needs, and refueling delays also trigger negative rewards. Successfully avoiding congestion and properly allocating resources provide additional positive incentives. Reward signals are calculated based on the actual results of each scheduling action and subsequent feedback to guide strategic adjustments.

[0055] The system continuously simulates various scheduling scenarios based on the current real-world environment, collects rewards and feedback from different scheduling actions, and optimizes policy network parameters using methods such as policy gradients. During actual operation, the system combines offline training with online fine-tuning to continuously improve the policy network's adaptability and global optimality. In special scenarios (such as equipment failures and emergency scheduling needs), it supports real-time online policy adjustments to rapidly respond to changes.

[0056] The reinforcement learning policy network outputs multiple scheduling solutions per round, encompassing different trade-offs (e.g., prioritizing energy consumption, timeliness, and task completion rate), all of which meet hard constraints. Each solution is scored and labeled with multiple performance criteria to facilitate subsequent Pareto screening and practical application selection.

[0057] The Pareto optimal solution under the current constraints is screened from the candidate set of scheduling schemes to generate a structured initial scheduling scheme; the initial scheduling scheme includes the start and stop timing of hydrogen production nodes, hydrogen storage container allocation, transportation route configuration and priority sorting of refueling tasks.

[0058] Specifically, the system conducts multi-objective comparisons on candidate solutions and automatically selects non-inferior solutions (i.e., Pareto optimal solutions) that are not "completely dominated" by other solutions under all optimization objectives to form the optimal solution domain.

[0059] Combined with current operational constraints and future demand changes, a structured initial scheduling plan is automatically generated, including: Start-up and shutdown sequence of hydrogen production nodes: clarify the start-up and shutdown time points of each hydrogen production device to ensure the balance between production and demand and optimal energy consumption.

[0060] Hydrogen storage container allocation: Rationally allocate the hydrogen inlet and outlet tasks of each storage tank to avoid the risks of over-storage and under-storage.

[0061] Transport route configuration: Assign specific vehicles, routes, start and end times to each transport batch or task to avoid congestion and high-risk sections.

[0062] Prioritization of refueling tasks: Prioritize the refueling tasks of each hydrogen refueling station or end user, and allocate resources first to key users or urgent tasks.

[0063] The hydrogen energy link scheduling optimization module is used to construct a hydrogen energy link scheduling model based on the initial scheduling plan using a graph structure path planning method based on an artificial potential field algorithm, dynamically avoid bottleneck nodes, congested sections, and high-risk areas in the transportation path, and generate an optimal scheduling instruction set for the entire link in combination with scheduling timing and resource cost constraints; The operation process of the hydrogen energy link scheduling optimization module includes the following steps: Based on the initial scheduling plan and in combination with the multi-dimensional operation data, a graph structure path planning method based on an artificial potential field algorithm is used to construct a hydrogen energy link scheduling model. The hydrogen energy link scheduling model uses hydrogen energy link nodes as vertices and transportation and transshipment paths as edges. Node attributes include resource capacity, task queue, and safety status. Edge attributes include path accessibility, latency, transportation risk level, and historical congestion records. Specifically, the system reads the initial scheduling plan and the latest multi-dimensional operating data, and automatically generates a global directed graph model of the hydrogen energy link: Node: represents hydrogen production plants, hydrogen storage stations, transportation hubs, refueling stations, end users, etc. Each node's attributes include real-time resource capacity (such as maximum / minimum available capacity), current task queue, and historical and current safety status (such as equipment health and accident frequency).

[0064] Edges: Represent hydrogen transportation and transshipment routes, such as pipelines, roads, and tracks. Each edge's attributes include path accessibility (whether it's unobstructed), expected latency, transportation risk level (such as natural disasters and traffic jams), and historical congestion records.

[0065] Synchronize the status information of each node and path in real time, integrate historical operation data and current monitoring data, and ensure that the model reflects the true dynamic status of the entire link.

[0066] Based on the hydrogen energy link scheduling model, attractive and repulsive potentials are applied to each transport route to identify and mark potential bottleneck sections and high-risk nodes in the route in real time; the attractive potential is used to guide the scheduled object to approach the target node, and the repulsive potential is used to dynamically avoid bottleneck nodes, congested sections and high-risk areas; Specifically, the system automatically assigns attractive potential to all target nodes (such as terminal demand points and key refueling stations). This attractive potential can be dynamically adjusted based on the urgency of the task, demand volume, and priority level of each node. For example, a point for emergency medical hydrogen use could be assigned a stronger attractive potential to ensure priority supply. The attractive potential parameters for target nodes can be updated in real time as scheduling objectives change, allowing for a temporary increase in the attractive potential of a specific node in the event of an emergency.

[0067] Repulsion Potential Assignment: The system dynamically assigns high repulsion potential to detected bottleneck nodes (e.g., a tank nearing capacity, a congested transport channel, or a piece of equipment experiencing frequent alarms), prompting the route planning process to avoid these risky areas. Repulsion potential parameters are calculated based on factors such as the node / route's real-time capacity, historical failures, current failure rate, and accident probability, and are automatically adjusted based on monitoring results.

[0068] The system continuously collects real-time operational data for each node / route, such as remaining capacity, queue length, task backlog, and equipment status. When a node reaches or approaches a set threshold (e.g., capacity utilization exceeding 90%, equipment health falling below the safety line, or failure frequency exceeding the historical average), it is automatically marked as a bottleneck node and its exclusion potential is updated. Historical data mining supports the early prediction of potential bottlenecks, such as by combining time series analysis to predict possible future backlogs on transport routes.

[0069] By combining multi-dimensional data (such as meteorological, traffic, security, and equipment warnings), certain areas are automatically identified as high-risk. If affected by extreme weather, road construction, or environmental anomalies, the relevant paths and nodes will be immediately adjusted for exclusion. For "known high-risk + temporary emergencies" events, the system automatically pushes risk updates, and the relevant nodes / paths are immediately implemented in the model.

[0070] Potential field parameters are refreshed in real time at fixed intervals (e.g., 5 minutes) or event-driven to ensure that the model reflects the latest changes in link status and implement dynamic security control.

[0071] Using a graph-structured path search method, under the influence of an artificial potential field, the transport path of each scheduling task is dynamically and iteratively adjusted, scheduling timing is optimized, bottleneck sections and high-risk nodes are proactively avoided, and the optimal scheduling plan for the entire link is generated based on the actual link resource status and scheduling task requirements. Specifically, within the graph model, the system utilizes an artificial potential field algorithm to perform path search for each scheduled task. The task's starting point actively searches for the target node based on attractive potential, automatically avoiding paths and nodes with high repulsive potential. Dynamic multi-path parallel search is supported, generating multiple alternative paths for the same task and evaluating them in real time, including timeliness, resource consumption, and risk.

[0072] Whenever the system receives new real-time operational data (such as sudden congestion, equipment anomalies, or changes in task status), it immediately triggers dynamic adjustments to local routes. If the primary route is deemed high-risk or unreachable, the system automatically selects a suboptimal or backup route, adjusting the scheduled task's departure time, route, or even switching transport vehicles. High-priority tasks are prioritized for optimal resources and routes, while low-priority tasks are automatically queued or have their shipping plans adjusted.

[0073] Path search results are connected to the full-link resource pool in real time, ensuring that each scheduling decision avoids over-allocation and local resource depletion. An optimal time window and resource allocation table are automatically generated for each task, properly avoiding peak periods, equipment maintenance periods, or other occupied resources. By comprehensively evaluating global timeliness, transportation and storage resource utilization, and risk distribution, the system automatically generates a full-link scheduling solution that meets all constraints and achieves optimal performance. The system possesses continuous self-learning capabilities, continuously adjusting path search preferences and resource allocation logic based on historical execution feedback to improve scheduling robustness and responsiveness.

[0074] The optimal scheduling scheme is encoded to generate a full-link optimal scheduling instruction set, which includes a task execution schedule, an allocation map of transportation and storage resources, a path adjustment and switching strategy, and a key node monitoring priority sorting.

[0075] Specifically, the optimized scheduling plan is automatically transcoded into a structured scheduling instruction set, which is convenient for the system downstream modules and on-site equipment to directly call and execute.

[0076] The instruction set includes: Task execution schedule: clearly define the planned start / end time points for each scheduled task.

[0077] Transportation and storage resource allocation map: Detailed allocation of resources such as vehicles, pipelines, and tanks used for each task.

[0078] Path adjustment and switching strategy: backup paths and emergency switching procedures when sudden congestion / risk changes occur.

[0079] Prioritization of key node monitoring: Dynamically generate a list of nodes and paths that require key monitoring to improve security and operation and maintenance response efficiency.

[0080] The instruction set is automatically pushed to the execution system and monitoring center, supporting real-time monitoring of scheduling execution effects, exception feedback and secondary optimization.

[0081] Furthermore, the formula of the hydrogen energy link scheduling model is as follows: in, Represents the value of the system's scheduling guidance function from node i to node j at time t, which is used to measure whether the task is currently prioritized to be transmitted through this path. The larger the value, the more suitable the path is for scheduling hydrogen tasks. This function will be directly used for path selection and task routing priority sorting in the graph structure, and is the input basis for generating the optimal scheduling instruction set for the entire link; represents the spatial gradient operator of the artificial potential field in the path network; represents the attractive potential function, which is used to measure the node resource adequacy and relay coordination degree; represents the repulsive potential function, which is used to measure path risks (such as equipment failure, environmental disturbance) and historical congestion; represents the disturbance term on the path; represents the current task load intensity of node j at time t; represents the maximum resource service capability of node j; represents the path scheduling coupling strength from source node i to target node j through relay node k; represents the current remaining available resource ratio of relay node k; M represents the total number of relay nodes that can participate in cooperative scheduling in the current path; represents the risk level of the path from node i to node j; Indicates the historical congestion level from node i to node j; and represents the weight coefficient of the attractive potential term; Represents the weight coefficient of the repulsive potential term.

[0082] The calculation formula is as follows: in,; represents the path scheduling coupling strength from source node i to target node j through relay node k; and represents the path distance from i to k and from k to j; Indicates the current task waiting queue length of relay node k.

[0083] The safety feedback module is used to collect real-time monitoring data based on the full-link optimal scheduling instruction set, use fuzzy logic control rules to perform safety assessment and intervention decisions on link operation, identify abnormal conditions and trigger closed-loop control feedback, and output safety intervention instructions, including scheduling correction suggestions, risk warning information and safety control parameter adjustment instructions.

[0084] It should be noted that the fuzzy logic control rules are specifically combined with typical risk scenarios of each node in the entire hydrogen energy chain (such as overpressure, leakage, overload, hysteresis, long-term high temperature, valve abnormality, etc.) to establish a multi-level fuzzy logic rule base: Input variables: pressure, temperature, flow rate, hydrogen concentration, transportation delay, equipment abnormality rate, alarm signal, etc. The membership function is adaptively adjusted and automatically optimized based on historical accident data; Rule examples: If the "tank pressure is high", the "temperature rises rapidly" and the "hydrogen leak signal is true", then the "safety risk is extremely high"; If the "filling flow rate is too high" and the "terminal load fluctuates", "speed limit adjustment is required".

[0085] Perform real-time fuzzy reasoning on the status of each node and the entire link, and output: security level classification (safe, concerned, warning, emergency); risk cause location (such as equipment aging, environmental deterioration, scheduling conflicts, etc.); decision threshold for triggering intervention mechanisms.

[0086] Combining fuzzy inference output and time series data, the following anomaly types can be automatically identified: storage and transportation overpressure, leakage, overtemperature, explosion criticality; transportation delays or route deviations; equipment health degradation (based on vibration spectrum, energy consumption anomalies, etc.); data islands / dead zones (signal loss).

[0087] Once the risk level reaches the preset threshold, the module automatically outputs the following safety intervention instructions: Scheduling correction suggestions: adjust transportation routes (such as switching to alternative routes), optimize scheduling timing (such as delaying / advancing refueling), and change the order of hydrogen storage tanks; Risk warning information: Detailed warning information will be pushed via SMS / APP / platform screen, indicating risk nodes, causes and recommended measures; Safety control parameter adjustment: automatically issue pressure / flow limits, start / stop commands, or initiate emergency plans (such as emergency shutdown, emptying, isolation, ventilation, etc.).

[0088] Example 2 A hydrogen energy full-link scheduling method based on big data, comprising the following steps: Through the distributed sensor network, multi-dimensional operation data of the entire hydrogen energy chain is collected in real time; Based on the multi-dimensional operating data, a hydrogen energy supply and demand forecasting model constructed by stochastic differential equations and neural stochastic processes is used to dynamically evolve and forecast the spatiotemporal dynamic change trends of hydrogen supply and demand, and output supply and demand forecast results; Based on the supply and demand forecast results and in combination with the multi-dimensional operation data, link operation constraints are extracted, and a multi-objective joint optimization strategy driven by reinforcement learning is adopted to generate an initial scheduling plan; Based on the initial scheduling plan, a graph-structured path planning method based on an artificial potential field algorithm is used to construct a hydrogen energy link scheduling model. Dynamic avoidance calculations are performed on bottleneck nodes, congested sections, and high-risk areas in the transportation path. Combined with scheduling timing and resource cost constraints, an optimal scheduling instruction set for the entire link is generated. Based on the full-link optimal scheduling instruction set, real-time monitoring data is collected, and fuzzy logic control rules are used to conduct safety assessment and intervention decisions on link operation, identify abnormal conditions and trigger closed-loop control feedback, and output safety intervention instructions.

[0089] In this embodiment, a hydrogen energy full-link scheduling method based on big data is applied to a hydrogen energy full-link scheduling system based on big data described in Example 1, which will not be repeated here.

[0090] In summary, the present invention has constructed a multi-dimensional operation data collection system with a distributed sensor network as the core, covering key indicators such as hydrogen production, purity, pressure, temperature, transportation route, terminal load, environmental parameters and equipment status. Combined with edge computing and 5G communication, it realizes full-link, low-latency data aggregation, providing high-quality support for prediction and decision-making.

[0091] The supply and demand forecasting module in the present invention introduces the joint modeling of stochastic differential equations and neural stochastic processes to accurately capture the dynamic coupling behavior and uncertain disturbances in the supply and demand system, output supply and demand trends, gap forecasts and confidence intervals, and achieve highly forward-looking control of the operating status of the hydrogen energy link. The model has the ability of adaptive structural adjustment and parameter calibration, and can maintain prediction stability under extreme events. The hydrogen production, storage and transportation matching module adopts a reinforcement learning optimization strategy, constructs a multi-objective joint optimization model based on link constraints and prediction results, and outputs an initial scheduling plan covering hydrogen production start and stop, storage tank allocation, transportation route and refueling sorting. The reinforcement learning network realizes a multi-dimensional trade-off between timeliness, energy consumption, completion rate and risk, and enhances the system's intelligent decision-making ability and global optimality.

[0092] The scheduling optimization module in this invention constructs a graph-structured path planning model through an artificial potential field algorithm, integrating attractive and repulsive potential mechanisms to achieve dynamic avoidance of bottleneck nodes and high-risk areas. Combining resource capacity and historical status, it generates an optimal scheduling instruction set for the entire link, improving the timeliness, flexibility, and safety of path planning. Based on fuzzy logic control rules, the safety feedback module integrates real-time data to classify link safety status and identify risks. It can also automatically output scheduling correction suggestions, risk warnings, and parameter intervention instructions, establishing a closed-loop safety response mechanism that effectively ensures the system's stable operation and risk prevention and control capabilities.

[0093] The above embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary engineering technicians in this field should fall within the scope of protection determined by the claims of the present invention.

Claims

1. A hydrogen energy full-link scheduling system based on big data, characterized in that: It includes a hydrogen energy link data acquisition module, a hydrogen energy supply and demand forecasting module, a hydrogen production, storage and transportation matching module, a hydrogen energy link scheduling optimization module and a safety feedback module which are sequentially connected in communication; The hydrogen energy link data acquisition module is used to collect multi-dimensional operating data of the entire hydrogen energy link in real time through a distributed sensor network; the entire link includes hydrogen production, compression, storage, transportation, filling and terminal utilization; The hydrogen energy supply and demand forecasting module is used to dynamically evolve and predict the spatiotemporal dynamic change trends of the hydrogen supply and demand sides based on the multi-dimensional operating data using a hydrogen energy supply and demand forecasting model constructed by stochastic differential equations and neural stochastic processes, and output supply and demand forecast results; The hydrogen production, storage and transportation matching module is used to extract link operation constraints based on the supply and demand forecast results and the multi-dimensional operation data, and generate an initial scheduling plan using a multi-objective joint optimization strategy driven by reinforcement learning; The hydrogen energy link scheduling optimization module is used to construct a hydrogen energy link scheduling model based on the initial scheduling plan using a graph structure path planning method based on an artificial potential field algorithm, dynamically avoid bottleneck nodes, congested sections, and high-risk areas in the transportation path, and generate an optimal scheduling instruction set for the entire link in combination with scheduling timing and resource cost constraints; The safety feedback module is used to collect real-time monitoring data based on the full-link optimal scheduling instruction set, use fuzzy logic control rules to perform safety assessment and intervention decisions on link operation, identify abnormal conditions and trigger closed-loop control feedback, and output safety intervention instructions, including scheduling correction suggestions, risk warning information and safety control parameter adjustment instructions.

2. A hydrogen energy full-link scheduling system based on big data according to claim 1, characterized in that: The multi-dimensional operating data includes hydrogen production, purity, pressure, temperature, remaining capacity of storage tanks, transportation route status, filling flow rate, terminal load demand, environmental parameters and equipment operating status.

3. A hydrogen energy full-link scheduling system based on big data according to claim 1, characterized in that: The operation process of the hydrogen energy supply and demand forecasting module includes the following steps: Performing preprocessing operations on the multidimensional operating data, including aggregation, performing anomaly removal, missing completion, and spatiotemporal feature normalization, and reconstructing the multidimensional operating data into a structured, time-series continuous input feature matrix; Based on the input feature matrix, a hydrogen supply and demand forecasting model is constructed by using stochastic differential equations and neural stochastic processes to characterize the dynamic evolution characteristics of hydrogen supply and demand; The input feature matrix is ​​input into the hydrogen energy supply and demand forecasting model, and numerical integration and neural network-driven random process deduction are used to dynamically simulate and predict the spatiotemporal dynamic change trends of the hydrogen supply and demand sides in each region, and output supply and demand forecast results, including hydrogen supply and demand curves, corresponding confidence intervals, supply and demand trends in each region in each future time period, supply and demand gaps, predicted confidence intervals and potential risk sections.

4. A hydrogen energy full-link scheduling system based on big data according to claim 3, characterized in that: The process of constructing the hydrogen energy supply and demand forecasting model includes the following steps: Based on the multi-dimensional operating data, a stochastic differential equation containing a state variable coupling term and a noise driving term is constructed to simulate the random disturbance characteristics and dynamic coupling behavior in the hydrogen supply and demand process; A neural stochastic process framework is introduced, and the solution space of the stochastic differential equation is used as the prior structure. Through joint training on historical supply and demand data and related environmental characteristics, the nonlinear time series distribution law of the supply and demand curves of each region is learned; Dynamically adjusting the kernel function parameters and prior distribution in the neural stochastic process using the maximum a posteriori estimation method, structurally coupling the stochastic differential equations with the neural stochastic process model to form a hydrogen energy supply and demand prediction model; Based on the hydrogen energy supply and demand forecasting model, the multi-dimensional operation data is input, a numerical solver is used to deduce the future trajectory of supply and demand distribution, and a supply and demand forecasting result is output.

5. The hydrogen energy full-link scheduling system based on big data according to claim 3 is characterized in that: The formula of the hydrogen energy supply and demand forecasting model is as follows: in, Represents the hydrogen supply and demand state function of region a The first-order derivative relative to time T, that is, the direction and intensity of changes in the hydrogen supply and demand balance trend in the region; represents the hydrogen supply and demand state function of region a at time T; T represents the time step; Indicates that region b at the historical moment Multi-dimensional operation data; Represents a nonlinear feature mapping function based on a neural network; The kernel function representing the time sensitivity of the weighted integral; represents the nonlinear coupling response function of region b to region a; The static weight coefficient representing the inter-regional coupling strength; represents the system disturbance term of region a at time t; represents the second-order derivative of the supply and demand status of region a over time; N represents the total number of regions or hydrogen energy link nodes participating in the joint modeling; Represents the hydrogen supply and demand state function of region a The second derivative with respect to time T, i.e. the acceleration of the change in supply and demand conditions; Indicates the dynamic response inertia coefficient of area a.

6. The hydrogen energy full-link scheduling system based on big data according to claim 1 is characterized in that: The operation process of the hydrogen production, storage and transportation matching module includes the following steps: Based on the multi-dimensional operation data and supply and demand forecast results, link operation constraints are extracted, including the production capacity boundary of the hydrogen production device, the capacity limit of the hydrogen storage container, the accessibility map of the transportation path, the service time window and priority level of the refueling node; The link operation constraints are combined to construct a joint optimization goal, which includes minimizing hydrogen production energy consumption, minimizing storage and transportation delays, maximizing task completion rate, and avoiding path congestion; Based on the joint optimization objective, a reinforcement learning framework is constructed that integrates a hierarchical state encoder and a multi-strategy behavior network. The dynamic supply and demand forecast results are used as state inputs, and the link resource scheduling cost and the supply and demand matching deviation are used as joint reward signals. The policy network is trained using a policy gradient optimization method to output a candidate set of scheduling solutions for the optimal solution domain. The Pareto optimal solution under the current constraints is screened from the candidate set of scheduling schemes to generate a structured initial scheduling scheme; the initial scheduling scheme includes the start and stop timing of hydrogen production nodes, hydrogen storage container allocation, transportation route configuration and priority sorting of refueling tasks.

7. The hydrogen energy full-link scheduling system based on big data according to claim 1 is characterized in that: The operation process of the hydrogen energy link scheduling optimization module includes the following steps: Based on the initial scheduling plan and in combination with the multi-dimensional operation data, a graph structure path planning method based on an artificial potential field algorithm is used to construct a hydrogen energy link scheduling model. The hydrogen energy link scheduling model uses hydrogen energy link nodes as vertices and transportation and transshipment paths as edges. Node attributes include resource capacity, task queue, and safety status. Edge attributes include path accessibility, latency, transportation risk level, and historical congestion records. Based on the hydrogen energy link scheduling model, attractive and repulsive potentials are applied to each transport route to identify and mark potential bottleneck sections and high-risk nodes in the route in real time; the attractive potential is used to guide the scheduled object to approach the target node, and the repulsive potential is used to dynamically avoid bottleneck nodes, congested sections and high-risk areas; Using a graph-structured path search method, under the influence of an artificial potential field, the transport path of each scheduling task is dynamically and iteratively adjusted, scheduling timing is optimized, bottleneck sections and high-risk nodes are proactively avoided, and the optimal scheduling plan for the entire link is generated based on the actual link resource status and scheduling task requirements. The optimal scheduling scheme is encoded to generate a full-link optimal scheduling instruction set, which includes a task execution schedule, an allocation map of transportation and storage resources, a path adjustment and switching strategy, and a key node monitoring priority sorting.

8. A hydrogen energy full-link scheduling system based on big data according to claim 7, characterized in that: The formula of the hydrogen energy link scheduling model is as follows: in, It represents the value of the system's scheduling guidance function from node i to node j at time t; represents the spatial gradient operator of the artificial potential field in the path network; represents the attractive potential function; represents the repulsive potential function; represents the disturbance term on the path; represents the current task load intensity of node j at time t; represents the maximum resource service capability of node j; represents the path scheduling coupling strength from source node i to target node j through relay node k; represents the current remaining available resource ratio of relay node k; M represents the total number of relay nodes that can participate in cooperative scheduling in the current path; represents the risk level of the path from node i to node j; Indicates the historical congestion level from node i to node j; and represents the weight coefficient of the attractive potential term; Represents the weight coefficient of the repulsive potential term.

9. A hydrogen energy full-link scheduling method based on big data, characterized in that: The method is applied to a hydrogen energy full-link scheduling system based on big data as described in any one of claims 1 to 8, comprising the following steps: Through the distributed sensor network, multi-dimensional operation data of the entire hydrogen energy chain is collected in real time; Based on the multi-dimensional operating data, a hydrogen energy supply and demand forecasting model constructed by stochastic differential equations and neural stochastic processes is used to dynamically evolve and forecast the spatiotemporal dynamic change trends of hydrogen supply and demand, and output supply and demand forecast results; Based on the supply and demand forecast results and in combination with the multi-dimensional operation data, link operation constraints are extracted, and a multi-objective joint optimization strategy driven by reinforcement learning is adopted to generate an initial scheduling plan; Based on the initial scheduling plan, a graph-structured path planning method based on an artificial potential field algorithm is used to construct a hydrogen energy link scheduling model. Dynamic avoidance calculations are performed on bottleneck nodes, congested sections, and high-risk areas in the transportation path. Combined with scheduling timing and resource cost constraints, an optimal scheduling instruction set for the entire link is generated. Based on the full-link optimal scheduling instruction set, real-time monitoring data is collected, and fuzzy logic control rules are used to conduct safety assessment and intervention decisions on link operation, identify abnormal conditions and trigger closed-loop control feedback, and output safety intervention instructions.

Images