Multi-agent-based source network load storage cooperative scheduling green electricity hydrogen production method and system

Abstract

The invention provides a multi-agent-based source network load storage cooperative scheduling green electricity hydrogen production method and system, and relates to the field of novel power system and green electricity hydrogen production operation optimization, and the method comprises the steps: carrying out the modeling of the source network load storage of a wind-solar hydrogen production green electricity park, and obtaining a park model; dividing the park model to set a day-ahead time scale and an intra-day time scale, abstractly dividing equipment in the park model into a plurality of agents, constructing a corresponding local optimization model for each agent, and coupling all the local optimization models through a preset coordination quantity to obtain a coordination model; obtaining day-ahead information, inputting the day-ahead information into the collaborative model to obtain a day-ahead collaborative scheduling scheme, constructing an intra-day rolling optimization time window, and taking the day-ahead collaborative scheduling scheme as a constraint to generate an intra-day rolling scheduling scheme; and executing the intra-day rolling scheduling scheme to obtain an operation result, evaluating the operation result to obtain an operation deviation, adaptively adjusting the coordination quantity based on the operation deviation until the operation deviation is lower than a preset value, and outputting the corresponding intra-day rolling scheduling scheme.

Description

Technical Field

[0001] This invention relates to the field of new power system and green electricity hydrogen production operation optimization technology, specifically to a method and system for green electricity hydrogen production based on multi-agent source-grid-load-storage collaborative scheduling. Background Technology

[0002] The operation and scheduling of existing wind-solar-hydrogen production and integrated power-hydrogen systems are typically based on the principles of economic dispatching of the power system and peak shaving and valley filling on the load side. On the one hand, time-of-use pricing or spot pricing is used to optimize power purchase and sale plans and energy storage charging and discharging strategies. On the other hand, water electrolysis hydrogen production units are treated as ordinary electricity loads, and their operating power is determined through empirical rules or centralized optimization models. Some studies incorporate wind power, photovoltaics, energy storage, and hydrogen production loads into a unified optimization model, using methods such as mixed integer programming to conduct joint dispatching on the day-ahead or intraday time scale, taking into account both economic efficiency and renewable energy consumption.

[0003] Existing solutions mostly construct optimization models on a single time scale, such as focusing on optimization only for day-ahead planning or intraday operation. Even though some studies propose two-stage energy management, the constraint transmission and deviation closure between the upper-level planning and the lower-level correction are still weak, lacking a unified and coordinated framework covering both day-ahead and intraday time scales. As a result, the day-ahead stage fails to reserve sufficient adjustment space for energy storage and hydrogen production loads, and intraday operation can only be passively corrected, making it difficult to balance economic efficiency and renewable energy consumption. Summary of the Invention

[0004] This invention addresses the problems existing in the prior art by providing a method and system for green electricity hydrogen production based on multi-agent source-grid-load-storage coordinated scheduling. It can coordinate the flexibility of wind and solar power output, electrochemical energy storage and hydrogen production load on both day-ahead and intraday time scales. At the same time, it has good multi-agent distributed coordination capabilities to achieve overall optimization of park comprehensive costs, renewable energy consumption rate and green hydrogen ratio in complex and uncertain environments.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A model of the power generation, grid, load, and storage system in a wind-solar-hydrogen-to-green power industrial park is obtained. The park model is divided into two time scales: day-ahead and intraday. The equipment in the park model is abstracted into several agents. A corresponding local optimization model is built for each agent. The collaborative model is obtained by coupling all local optimization models through a preset coordination quantity. Obtain the previous day's information and input it into the collaborative model to obtain the previous day's collaborative scheduling scheme. Construct the intraday rolling optimization time window and use the previous day's collaborative scheduling scheme as a constraint to generate the intraday rolling scheduling scheme. The intraday rolling scheduling scheme is executed to obtain the running results. The running results are evaluated to obtain the running deviation. The coordination amount is adaptively adjusted based on the running deviation until the running deviation is lower than the preset value, and the corresponding intraday rolling scheduling scheme is output.

[0006] In some embodiments, the step of modeling the source-grid-load-storage of the wind-solar-hydrogen-green power park to obtain the park model includes: Construct a distribution network topology for the upstream grid interface points and key nodes in the wind-solar-hydrogen-green power industrial park; Modeling is performed using the distribution network topology as the network equipment, photovoltaic power plants and wind farms as the source equipment, electrochemical energy storage devices as the storage equipment, and water electrolysis for hydrogen production as the load equipment. The parameters between the wind-solar hydrogen production green power park and the upstream power grid, the hydrogen production load parameters of the water electrolysis hydrogen production device, and the energy storage parameters of the electrochemical energy storage device are configured to obtain the park model.

[0007] In some embodiments, the step of abstracting and dividing the equipment in the park model into several agents, constructing a corresponding local optimization model for each agent, and coupling all local optimization models through a preset coordination quantity to obtain a collaborative model includes: The equipment in the park model is abstracted and divided into several agents. Among them, the photovoltaic power station and wind farm are source agents, the park distribution network and its interface with the upper-level power grid are grid agents, the water electrolysis hydrogen production device is hydrogen production load agent, and the electrochemical energy storage device is energy storage agent. Build local optimization models for each agent; The optimization objective of the source agent is to calculate the decision between the utilization power of renewable energy and the curtailed power under a given coordination quantity; The optimization objective of the hydrogen production load proxy is to calculate the hydrogen production power decision for each time period; The optimization objective of the energy storage agent is to calculate the charging and discharging power decision within the SOC safe range; The optimization objective of the network agent is to calculate the power balance constraints of each agent; A collaborative model is obtained by coupling the power decisions of each agent with the power balance constraints of the network agent through a preset coordination quantity.

[0008] In some embodiments, the step of obtaining day-ahead information and inputting it into the cooperative model to obtain a day-ahead cooperative scheduling scheme includes: Obtain day-ahead information, which includes next-day photovoltaic and wind power output forecasts, park load, electricity / carbon price, hydrogen production demand, and green hydrogen ratio targets; Input the day-ahead information into the collaborative model, and the network agent generates an initial day-ahead plan based on the day-ahead information; Based on the preset coordination quantity, the source agent, hydrogen production load agent and energy storage agent in the collaborative model solve the local optimization problem respectively, and the network agent summarizes the results of each agent and updates the coordination quantity. Repeat the local optimization and update the coordination quantity until convergence, and use the day-ahead plan at the time of convergence as the day-ahead coordination scheduling scheme. The day-ahead coordinated scheduling scheme includes, but is not limited to, the power output curves of wind and solar power, the power purchased and sold, the SOC curve of energy storage, the power output curve of hydrogen production load, and the amount of wind and solar curtailment.

[0009] In some embodiments, the step of constructing an intraday rolling optimization time window and generating an intraday rolling scheduling scheme using the previous day's collaborative scheduling scheme as a constraint includes: Construct a rolling, optimized time window covering the next four hours of the day; The aforementioned day-ahead coordinated scheduling scheme is used as a constraint; Within the intraday rolling optimization time window, the collaborative model is re-controlled for local optimization to generate an intraday rolling scheduling scheme.

[0010] In some embodiments, the step of executing an intraday rolling scheduling scheme to obtain operating results, evaluating the operating results to obtain operating deviations, and adaptively adjusting the coordination amount based on the operating deviations includes: The results of executing the intraday rolling scheduling plan were obtained. Using electricity purchase and sale costs, wind and solar curtailment, green hydrogen ratio, unit hydrogen production electricity consumption and carbon emissions as indicators, the operational deviation is evaluated by the operational results. Based on operational deviations, the coordination quantity is adaptively adjusted. Record scheduling results and corresponding intraday rolling scheduling schemes, and build a strategy library.

[0011] In some embodiments, the optimization objective of the park model is: ; in, For electricity purchase price, For the power purchase capacity, For electricity sales price, For electricity sales capacity, This is the energy storage depreciation cost coefficient. For energy storage charging power, For energy storage discharge power, The penalty coefficient for wind and solar power curtailment. This refers to the power of wind and solar power that has been curtailed.

[0012] This invention proposes a system for green electricity-to-hydrogen production based on multi-agent source-grid-load-storage coordinated scheduling, comprising: The global unit is configured to model the source, grid, load, and storage of the wind-solar-hydrogen-green power park to obtain the park model; The local unit is configured to divide the park model into two time scales: day-ahead and intraday. The equipment in the park model is abstracted into several agents, and a corresponding local optimization model is built for each agent. A collaborative model is obtained by coupling all local optimization models through a preset coordination quantity. The scheduling scheme unit is configured to acquire day-ahead information, input it into the collaborative model, obtain the day-ahead collaborative scheduling scheme, construct the intraday rolling optimization time window, and use the day-ahead collaborative scheduling scheme as a constraint to generate the intraday rolling scheduling scheme. The adaptive unit is configured to execute an intraday rolling scheduling scheme to obtain the running results, evaluate the running results to obtain the running deviation, adaptively adjust the coordination amount based on the running deviation until the running deviation is lower than a preset value, and output the corresponding intraday rolling scheduling scheme.

[0013] This invention proposes a computer device, comprising: At least one processor; and a memory storing a computer program that can run on the processor, wherein the processor executes the program to perform the steps of the method and system for green electricity production based on multi-agent source-grid-load-storage coordinated scheduling.

[0014] This invention proposes a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the method and system for green electricity production based on multi-agent source-grid-load-storage collaborative scheduling.

[0015] Compared with the prior art, the present invention has the following beneficial effects: This invention proposes a method and system for green electricity-to-hydrogen production based on multi-agent source-grid-load-storage collaborative scheduling. The method includes: modeling the source, grid, load, and storage of a wind-solar-hydrogen-to-hydrogen green power park to obtain a park model; dividing the park model into two time scales, day-ahead and intraday; abstracting the equipment in the park model into several agents; constructing a corresponding local optimization model for each agent; coupling all local optimization models through a preset coordination quantity to obtain a collaborative model; acquiring day-ahead information and inputting it into the collaborative model to obtain a day-ahead collaborative scheduling scheme; constructing an intraday rolling optimization time window; using the day-ahead collaborative scheduling scheme as a constraint to generate an intraday rolling scheduling scheme; executing the intraday rolling scheduling scheme to obtain the running results; evaluating the running results to obtain the running deviation; adaptively adjusting the coordination quantity based on the running deviation until the running deviation is lower than a preset value; and outputting the corresponding intraday rolling scheduling scheme.

[0016] This invention enables integrated collaborative scheduling across two time scales for green power parks using wind, solar, and hydrogen production. It constructs a hierarchical collaborative scheduling framework covering both day-ahead and intraday time scales. Through boundary condition transmission and deviation feedback mechanisms, it organically connects power purchase and sale plans, energy storage SOC trajectories, and hydrogen production load operation plans between day-ahead and intraday time scales. This avoids the disconnect between planning and execution caused by traditional single-time-scale or loosely coupled two-stage scheduling, thereby improving the executability and operational economy of the scheduling scheme.

[0017] This invention explicitly models hydrogen production load as a flexible resource and optimizes it in conjunction with energy storage. It transforms hydrogen production load from a rigid load into a flexible resource similar to energy storage, explicitly modeling its adjustable power range, peak shifting window, cumulative hydrogen production, green hydrogen ratio, and process constraints. It also solves the SOC and charge / discharge power constraints of electrochemical energy storage in a unified manner, enabling hydrogen production load and energy storage to form a complementary flexible resource. This actively responds to wind and solar power output and electricity price signals, thereby improving the local consumption rate of renewable energy and the proportion of green hydrogen production.

[0018] This invention, based on distributed collaborative optimization involving multiple agents, enhances scalability and engineering adaptability. By partitioning the problem into multiple agents—source agent, network agent, hydrogen production load agent, and energy storage agent—it breaks down the centralized, large-scale optimization problem into several local sub-problems, and achieves global collaboration through iterative coordination. The architecture reduces solution complexity, adapts to day-ahead and intraday rolling scheduling requirements, and supports the partial disclosure of equipment-side models and data, thus facilitating data privacy protection among different entities and modular system expansion.

[0019] This invention comprehensively considers economic efficiency, green hydrogen ratio, and carbon reduction targets, and supports low-carbon operation optimization. The invention integrates electricity purchase and sale costs, energy storage depreciation costs, wind and solar curtailment penalties, green hydrogen ratio, and carbon emission indicators into the objective function, and can adaptively adjust the weights according to operational deviations. Under the premise of ensuring the safe operation of the park, it takes into account both economic efficiency and low-carbon targets, and adapts to the evaluation needs of green hydrogen certification and carbon emission constraints.

[0020] This invention features an adaptive scheduling strategy with superior long-term performance. Through operational deviation evaluation and strategy library maintenance, this invention can automatically adjust target weights, constraint boundaries, and coordination algorithm parameters based on historical operational results. This allows the scheduling strategy to gradually adapt to changes in wind and solar resources, electricity pricing mechanisms, and load characteristics during long-term operation, thereby improving the robustness and engineering practicality of the scheduling strategy.

[0021] This invention's collaborative scheduling enables optimized allocation of green electricity over a wider area. Intelligent agents can transmit surplus green electricity from areas with surplus resources to areas with insufficient power for hydrogen production, based on the distribution of green electricity resources and hydrogen production needs in different regions, thus improving the overall absorption capacity of green electricity. As the scale of green electricity-to-hydrogen production expands, dependence on traditional fossil fuels will decrease, carbon emissions will be reduced, and a clean and low-carbon energy transition will be facilitated. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these drawings without creative effort.

[0023] Figure 1 The flowchart of the green electricity hydrogen production method based on multi-agent source-grid-load-storage collaborative scheduling provided by the present invention is shown.

[0024] Figure 2 The system module diagram for green electricity-to-hydrogen production based on multi-agent source-grid-load-storage collaborative scheduling provided by the present invention.

[0025] Figure 3 A schematic diagram of the structure of an embodiment of the computer device provided by the present invention.

[0026] Figure 4 This is a schematic diagram of an embodiment of the computer-readable storage medium provided by the present invention.

[0027] Figure 5 This is a schematic diagram of the source-grid-load-storage structure of a wind-solar hydrogen production green power park, as provided in the embodiment of the multi-agent source-grid-load-storage collaborative scheduling method for green electricity production based on the present invention.

[0028] Figure 6 This is a schematic diagram of a multi-timescale source-grid-load-storage collaborative scheduling framework for green electricity production based on multi-agent source-grid-load-storage collaborative scheduling in a wind-solar hydrogen production green power park, as provided by the present invention.

[0029] Figure 7 This is a schematic diagram of a distributed collaborative optimization structure of source, grid, hydrogen production load and energy storage based on multi-agent source-grid-load-storage coordinated scheduling method for green electricity-to-hydrogen production provided by the present invention.

[0030] Figure 8 This is a flowchart illustrating the distributed collaborative optimization scheduling method of source-grid-load-storage in an embodiment of the source-grid-load-storage-based green electricity-to-hydrogen method provided by the present invention. Detailed Implementation

[0031] The present invention will now be further described with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and should not be construed as limiting the scope of protection of the present invention. It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of this application.

[0032] It should be noted that all uses of "first" and "second" in the embodiments of the present invention are for the purpose of distinguishing two entities with the same name but different names or different parameters. It can be seen that "first" and "second" are only for the convenience of expression and should not be construed as limiting the embodiments of the present invention. Subsequent embodiments will not explain this in detail.

[0033] With the proposal of dual-carbon targets and the rapid development of large-scale wind power, photovoltaics, and other renewable energy sources, green power parks that utilize a high proportion of renewable energy to produce hydrogen are gradually becoming an important development direction. In these parks, by configuring various equipment such as photovoltaic power plants, wind farms, electrochemical energy storage systems, and water electrolysis hydrogen production devices, integrated operation of source, grid, load, and storage can be achieved, improving the local consumption level of renewable energy and reducing carbon emissions from end-use energy.

[0034] The operation and scheduling of existing wind-solar-hydrogen production and integrated power-hydrogen systems are typically based on the principles of economic dispatching of the power system and peak shaving and valley filling on the load side. On the one hand, time-of-use pricing or spot pricing is used to optimize power purchase and sale plans and energy storage charging and discharging strategies. On the other hand, water electrolysis hydrogen production units are treated as ordinary electricity loads, and their operating power is determined through empirical rules or centralized optimization models. Some studies incorporate wind power, photovoltaics, energy storage, and hydrogen production loads into a unified optimization model, using methods such as mixed integer programming to conduct joint dispatching on the day-ahead or intraday time scale, taking into account both economic efficiency and renewable energy consumption.

[0035] However, from the perspective of integrated operation of power generation, grid, load, and storage in wind-solar-hydrogen production green power parks, existing technologies still have the following shortcomings. The lack of coordination across multiple time scales is a significant problem. Existing solutions often build optimization models on a single time scale, such as focusing on day-ahead planning or intraday operations. Even with some studies proposing two-stage energy management, the constraint transfer and deviation closure between upper-level planning and lower-level adjustments remain weak, lacking a unified coordination framework covering both day-ahead and intraday time scales. As a result, sufficient adjustment space for energy storage and hydrogen production loads is not reserved during the day-ahead phase, and intraday operations can only be passively adjusted, making it difficult to balance economic efficiency with renewable energy consumption.

[0036] The flexibility of hydrogen production load is not fully utilized. Current work often simplifies water electrolysis hydrogen production units into electrical loads that can be adjusted within a certain range, only considering constraints such as power limits and total hydrogen production. To ensure the solvability of the problem, fixed efficiency or piecewise linear models are often used. The time-shifting capacity of hydrogen production load, operating window constraints, the impact of start-up and shutdown frequency on lifespan, and the proportion of green hydrogen are difficult to systematically characterize and utilize, resulting in conservative scheduling strategies. The potential of hydrogen production load in absorbing wind and solar power fluctuations and increasing the proportion of green hydrogen is not fully realized.

[0037] Centralized scheduling suffers from insufficient scalability and adaptability. Most integrated source-grid-load-storage scheduling models employ a centralized modeling approach where a single scheduling center controls all equipment parameters and statuses, constructing a large-scale mixed-integer optimization problem. As the scale of the industrial park expands and the types of equipment increase, the model dimensionality and solution complexity rise sharply, making it difficult to meet the needs of intraday rolling optimization and hindering data privacy protection among different entities and modular expansion of the system. Some works have attempted to introduce decomposition coordination or heuristic algorithms, but overall, the approach remains primarily centralized model + decomposition solution, lacking a multi-agent collaborative scheduling mechanism oriented towards the autonomous characteristics of source, grid, load, and storage.

[0038] Economic efficiency, green hydrogen ratio, and carbon reduction targets are not closely coupled. Existing methods mostly focus on minimizing operating costs and lack clear modeling and constraints on indicators such as renewable energy integration rate, green hydrogen production ratio, and carbon emissions per unit of hydrogen production. These indicators are often only reflected through additional penalty items or empirical coefficients, which is difficult to meet the new evaluation needs such as green hydrogen certification and carbon emission assessment.

[0039] In summary, existing technologies lack a source-grid-load-storage coordinated scheduling method and system specifically designed for wind-solar-hydrogen-green power parks. This system should be able to coordinate the flexibility of wind and solar power output, electrochemical energy storage, and hydrogen production load on both day-ahead and intraday timescales, while also possessing good multi-agent distributed coordination capabilities. This would enable the overall optimization of the park's comprehensive costs, renewable energy absorption rate, and green hydrogen ratio under complex and uncertain environments.

[0040] The purpose of this invention is to address the problems existing in the source-grid-load-storage coordination of wind-solar-hydrogen-green power parks, such as insufficient coordination across multiple time scales, insufficient flexible utilization of hydrogen production load, poor scalability of centralized scheduling, and weak coupling between economic efficiency and low-carbon goals. This invention proposes a multi-agent-based collaborative scheduling method and system for source-grid-load-storage in wind-solar-hydrogen-green power parks.

[0041] By constructing a hierarchical scheduling framework covering both day-ahead and intraday timescales, and introducing a multi-agent collaborative optimization mechanism involving source agents, grid agents, hydrogen production load agents, and energy storage agents, hydrogen production load and electrochemical energy storage are explicitly modeled as schedulable flexible resources. Taking into account indicators such as electricity purchase and sale costs, renewable energy consumption rate, and green hydrogen ratio, the wind-solar-hydrogen-green power park can achieve economic, safe, and low-carbon operation under uncertain operating conditions.

[0042] This invention proposes a method and system for green electricity-to-hydrogen production based on multi-agent source-grid-load-storage collaborative scheduling. Please refer to [link / reference]. Figure 1 ,include: S1. Model the source, grid, load, and storage of the wind-solar-hydrogen-green power park to obtain the park model; S2. Divide the park model into two time scales: day-ahead and intraday. Abstract the equipment in the park model into several agents. Build a corresponding local optimization model for each agent. Couple all local optimization models through a preset coordination quantity to obtain a collaborative model. S3. Obtain the previous day's information and input it into the collaborative model to obtain the previous day's collaborative scheduling scheme. Construct the intraday rolling optimization time window and use the previous day's collaborative scheduling scheme as a constraint to generate the intraday rolling scheduling scheme. S4. Execute the intraday rolling scheduling scheme to obtain the running results, evaluate the running results to obtain the running deviation, and adaptively adjust the coordination amount based on the running deviation until the running deviation is lower than the preset value, and output the corresponding intraday rolling scheduling scheme.

[0043] This invention addresses the needs of green energy parks that utilize wind and solar power for hydrogen production, proposing a source-grid-load-storage scheduling method based on multi-agent distributed collaboration. The method operates on both day-ahead and intraday timescales. The day-ahead layer optimizes 96 time periods from 00:15 to 24:00 the next day with a resolution of 15 minutes, generating wind power output plans, photovoltaic power output plans, energy storage charging and discharging plans, and electrolyzer power consumption plans. The intraday layer optimizes 16 time periods over the next 4 hours with a 15-minute time resolution. Based on real-time updated forecast data and equipment status, the day-ahead plan is rolled over and corrected every 15 minutes to achieve dynamic coordination of power generation and consumption plans.

[0044] Green electricity is intermittent and fluctuates, posing a challenge to the stable operation of hydrogen production systems. Multi-agent collaborative scheduling can effectively mitigate these fluctuations through energy storage system charging and discharging regulation, grid interaction, and complementarity between different energy sources. When photovoltaic or wind power output suddenly decreases, the energy storage system can rapidly release electricity to ensure the continuous operation of the electrolyzer; simultaneously, the agents obtain power from the grid to ensure the stability of the hydrogen production process. Through multi-agent partitioning, photovoltaic power plants, wind farms, electrochemical energy storage, hydrogen production loads, and the park's distribution network are abstracted as source agents, energy storage agents, hydrogen production load agents, and grid agents, respectively. Distributed collaborative optimization is performed under the guidance of coordination variables such as node equivalent electricity prices and power quotas. Real-time monitoring of grid electricity prices allows for the purchase of electricity from the grid or an increase in the proportion of green electricity used during periods of lower electricity prices, based on green electricity output and hydrogen production demand.

[0045] To model the power distribution network topology of the industrial park, real-time data from photovoltaic arrays, wind turbines, electrolyzers, energy storage devices, and grid connection points needs to be collected to construct a digital model that includes electrical parameters, equipment efficiency, and constraints. The model must reflect the intermittent characteristics of wind and solar power output, with photovoltaic output varying non-linearly with light intensity, and wind power being proportional to the cube of wind speed. Simultaneously, the power adjustment range of the electrolyzers and the charging and discharging efficiency of the energy storage devices must be considered.

[0046] Regarding multi-timescale segmentation, day-ahead optimization uses a 24-hour cycle, formulating equipment start-up and shutdown plans based on weather forecast data to determine the electrolyzer operating power during the next day's photovoltaic surplus period. Intraday optimization uses a 15-minute rolling window, dynamically adjusting scheduling commands by monitoring wind and solar power output deviations in real time. Agent segmentation requires decoupling equipment by function; for example, photovoltaic arrays, wind farms, and electrolyzer clusters are treated as independent agents, with each agent exchanging information through coordination quantities such as power commands and hydrogen production. The collaborative model construction adopts a distributed optimization framework, enabling each agent to minimize total cost while satisfying local constraints.

[0047] In some embodiments, please refer to Figure 1 The steps for modeling the power generation, grid, load, and storage systems of the wind-solar-hydrogen-to-green power park to obtain the park model include: Construct a distribution network topology for the upstream grid interface points and key nodes in the wind-solar-hydrogen-green power industrial park; Modeling is performed using the distribution network topology as the network equipment, photovoltaic power plants and wind farms as the source equipment, electrochemical energy storage devices as the storage equipment, and water electrolysis for hydrogen production as the load equipment. The parameters between the wind-solar hydrogen production green power park and the upstream power grid, the hydrogen production load parameters of the water electrolysis hydrogen production device, and the energy storage parameters of the electrochemical energy storage device are configured to obtain the park model.

[0048] The physical connection points between the industrial park and the upper-level power grid are clearly defined, enabling bidirectional power flow through a 110kV substation. The grid connection points of photovoltaic arrays, wind farms, electrolytic cell clusters, and energy storage devices are also identified, forming a topology including feeders, buses, and circuit breakers. Reflecting the hierarchical relationship of different voltage levels, the lines from the photovoltaic inverter outlet to the 35kV bus are considered primary branches, and the lines from the wind farm to the same bus are considered secondary branches, ensuring the accuracy of power flow direction and short-circuit current distribution.

[0049] The photovoltaic power plant model needs to incorporate the influence coefficients of light intensity and temperature on output power, and use a single diode equivalent circuit model to describe the IV characteristics of photovoltaic modules; the wind farm model needs to include wind speed power curves and pitch angle control logic, and use Weibull distribution to simulate the impact of wind speed fluctuations on unit output; the electrolyzer model needs to establish a nonlinear relationship between electro-hydrogen conversion efficiency and current density, and under rated operating conditions, each cubic meter of hydrogen production requires 4.5 kWh of electrical energy; the energy storage device model needs to define charge and discharge efficiency, upper and lower limits of SOC (state of charge), and self-discharge rate, with the charge and discharge efficiency of lithium-ion batteries set at 95% and the SOC range limited to 20%-90%.

[0050] The parameter configuration process requires the integration of multi-source data. Grid interface parameters include transformer capacity, line impedance, and power factor limits. Hydrogen production load parameters cover the rated power of the electrolyzer, ramp rate, and minimum operating time. The power adjustment range of a single electrolyzer is required to be 20%-100%, and the start-stop interval is not less than 30 minutes. Energy storage parameters involve capacity configuration, power response speed, and cycle life. A 2MWh energy storage system is configured to mitigate power fluctuations at the 1-hour level.

[0051] In some embodiments, please refer to Figure 1 The steps of abstracting and dividing the equipment in the park model into several agents, constructing a corresponding local optimization model for each agent, and coupling all local optimization models through a preset coordination quantity to obtain a collaborative model include: The equipment in the park model is abstracted and divided into several agents. Among them, the photovoltaic power station and wind farm are source agents, the park distribution network and its interface with the upper-level power grid are grid agents, the water electrolysis hydrogen production device is hydrogen production load agent, and the electrochemical energy storage device is energy storage agent. Build local optimization models for each agent; The optimization objective of the source agent is to calculate the decision between the utilization power of renewable energy and the curtailed power under a given coordination quantity; The optimization objective of the hydrogen production load proxy is to calculate the hydrogen production power decision for each time period; The optimization objective of the energy storage agent is to calculate the charging and discharging power decision within the SOC safe range; The optimization objective of the network agent is to calculate the power balance constraints of each agent; A collaborative model is obtained by coupling the power decisions of each agent with the power balance constraints of the network agent through a preset coordination quantity.

[0052] Photovoltaic and wind power generation outsourcing needs to address the intermittency and uncertainty of renewable energy. Photovoltaic outsourcing needs to adjust power allocation in response to upper-level coordination, while meeting the maximum power point tracking (MPPT) constraints of the inverter, based on real-time solar intensity forecasts. Wind power outsourcing needs to combine wind speed forecasts and pitch angle control logic to find the optimal solution between wind curtailment costs and power regulation flexibility. Hydrogen production load outsourcing needs to transform the electricity-hydrogen conversion process into a power decision problem. Electrolyzer clusters need to dynamically optimize power allocation for different time periods based on hydrogen demand forecasts, equipment start-up and shutdown constraints, and electricity price signals, while ensuring that process parameters such as hydrogen purity and pressure meet standards.

[0053] The optimization of energy storage agents needs to balance power and energy requirements. Lithium-ion batteries need to be kept within the safe SOC range of 20%-90%. Based on the charging and discharging commands determined by the coordinated quantity, the power response speed should be optimized to smooth out power fluctuations at the 1-hour level, while avoiding lifespan degradation caused by frequent charging and discharging. The grid agent needs to integrate the power decisions of all agents and verify whether the voltage of each node and the current carrying capacity of the line exceed the limit through power flow calculation. When a sudden increase in the power of the electrolyzer causes the 35kV bus voltage to drop by more than 5%, the grid agent needs to restore voltage stability by adjusting the energy storage charging and discharging strategy or coordinating the source agent to reduce its output.

[0054] In some embodiments, please refer to Figure 1 The steps of obtaining day-ahead information and inputting it into the cooperative model to obtain the day-ahead cooperative scheduling scheme include: Obtain day-ahead information, which includes next-day photovoltaic and wind power output forecasts, park load, electricity / carbon price, hydrogen production demand, and green hydrogen ratio targets; Input the day-ahead information into the collaborative model, and the network agent generates an initial day-ahead plan based on the day-ahead information; Based on the preset coordination quantity, the source agent, hydrogen production load agent and energy storage agent in the collaborative model solve the local optimization problem respectively, and the network agent summarizes the results of each agent and updates the coordination quantity. Repeat the local optimization and update the coordination quantity until convergence, and use the day-ahead plan at the time of convergence as the day-ahead coordination scheduling scheme. The day-ahead coordinated scheduling scheme includes, but is not limited to, the power output curves of wind and solar power, the power purchased and sold, the SOC curve of energy storage, the power output curve of hydrogen production load, and the amount of wind and solar curtailment.

[0055] To obtain the output forecast curves of photovoltaic and wind power at 15-minute intervals for the next day, photovoltaic forecasts need to be combined with historical radiation data and cloud cover models, while wind power forecasts need to consider the impact of terrain on wind speed distribution. At the same time, it is necessary to collect the electricity demand of non-hydrogen production loads in the park, the grid time-of-use electricity price and carbon trading price signals, as well as the production targets and green hydrogen ratio requirements stipulated in the green hydrogen contract.

[0056] Based on the input data, an initial day-ahead plan is generated, assuming full-power operation of photovoltaics and wind power output at maximum dispatchable capacity. Simultaneously, the charging and discharging periods for energy storage are planned according to the electricity price curve. The source agents (PV and wind power) solve local optimization problems under given coordination variables. The PV agent needs to minimize curtailment costs while satisfying the inverter's maximum power point tracking constraint; the wind power agent needs to balance pitch angle adjustment losses and power flexibility. The hydrogen production load agent optimizes power allocation for each time period based on hydrogen demand and electricity price signals, increasing electrolyzer power during periods of low electricity prices to reduce production costs; the energy storage agent plans charging and discharging strategies within the safe SOC range to smooth out power fluctuations.

[0057] After the grid agent aggregates the local solutions from each agent, the global feasibility is verified through power balance constraints. If line overload or voltage exceeding limits is detected, the coordination quantity is adjusted, triggering a new round of local optimization. Convergence is accelerated by the Alternating Directional Multiplier Method (ADMM), typically reaching the preset accuracy within 5-10 iterations, with a power deviation of less than 1%. The final day-ahead coordinated scheduling scheme includes several key curves: the wind and solar power output curve shows strong solar power and weak wind power during the day, and strong wind power and no solar power output at night; the power purchase and sale curve clarifies the interaction time and power with the grid; the energy storage SOC curve plans the daily charging and discharging rhythm; the hydrogen production load power curve ensures that hydrogen production meets the target; and the wind and solar curtailment curve quantifies the utilization rate of renewable energy.

[0058] Multi-agent systems can monitor the output of green electricity such as solar and wind power in real time, as well as the hydrogen production demand of electrolyzers. Through the analysis and processing of large amounts of data, the agents can accurately predict fluctuations in green electricity and adjust the operating status of each link in the power generation, grid, load, and storage system in advance, achieving a precise match between green electricity supply and hydrogen production demand, and reducing the waste of green electricity. When it is predicted that solar power output will increase in the near future, the agents can adjust the charging strategy of the energy storage system in advance to store excess green electricity for use in hydrogen production.

[0059] In some embodiments, please refer to Figure 1 The step of constructing an intraday rolling optimization time window and using the previous day's collaborative scheduling scheme as a constraint to generate an intraday rolling scheduling scheme includes: Construct a rolling, optimized time window covering the next four hours of the day; The aforementioned day-ahead coordinated scheduling scheme is used as a constraint; Within the intraday rolling optimization time window, the collaborative model is re-controlled for local optimization to generate an intraday rolling scheduling scheme.

[0060] Starting from the current moment, a rolling optimization window covering the next 4 hours is constructed. For example, at 10:00, an optimization interval of 10:15-14:15 is generated. The window scrolls forward every 15 minutes as time progresses, forming a continuous decision sequence.

[0061] Within the rolling window, the power generation agent adjusts its output forecast based on real-time meteorological data, changing the photovoltaic forecast from the planned 4MW to the actual 3.5MW; the hydrogen production agent dynamically adjusts power commands based on hydrogen inventory and production rate; and the energy storage agent optimizes charging and discharging strategies based on SOC status and power deviation. The grid agent aggregates the decisions from all agents and verifies line current carrying capacity and voltage deviation through power flow calculations. If any deviations are found, an adjustment signal is sent back. When a sudden increase in wind power output causes the 35kV bus voltage to rise to 1.05pu, the grid agent will request the energy storage to increase its charging capacity to absorb the excess power.

[0062] In some embodiments, please refer to Figure 1 The steps of executing the intraday rolling scheduling scheme to obtain operating results, evaluating the operating results to obtain operating deviations, and adaptively adjusting the coordination quantity based on the operating deviations include: The results of executing the intraday rolling scheduling plan were obtained. Using electricity purchase and sale costs, wind and solar curtailment, green hydrogen ratio, unit hydrogen production electricity consumption and carbon emissions as indicators, the operational deviation is evaluated by the operational results. Based on operational deviations, the coordination quantity is adaptively adjusted. Record scheduling results and corresponding intraday rolling scheduling schemes, and build a strategy library.

[0063] The intraday rolling scheduling scheme is implemented, and the optimized power command is sent to each equipment agent. The actual operation data is recorded at a frequency of minutes, including key parameters such as grid interaction power, energy storage SOC changes, and hydrogen production.

[0064] The evaluation of operational results requires the construction of a multi-dimensional indicator system. The cost of electricity purchase and sale is calculated by weighting the actual purchased electricity volume with time-of-use pricing; purchasing more electricity during the midday low-price period can reduce the total cost. The amount of wind and solar power curtailment is obtained by comparing actual output with the maximum available power of the equipment, reflecting the utilization rate of renewable energy. The green hydrogen ratio needs to verify the proportion of renewable energy electricity in the hydrogen source to ensure compliance with contract requirements. The unit hydrogen production electricity consumption is calculated as the ratio of total electricity consumption to hydrogen production, measuring electrolysis efficiency. Carbon emissions are estimated based on the product of the grid carbon emission factor and the purchased electricity volume, reflecting environmental benefits. Taking a certain industrial park as an example, if the actual cost of electricity purchase and sale is 8% higher than the current plan, the wind curtailment rate reaches 5%, and the green hydrogen ratio is only 92%, then deviation analysis is needed to pinpoint the root cause of the problem. This could be due to a deviation in photovoltaic forecasting leading to increased electricity purchase, or a sudden increase in wind power causing wind curtailment.

[0065] The adaptive adjustment of the coordination quantity adopts a proportional-integral-derivative (PID) control strategy. The proportional term responds quickly to the current deviation; when the wind curtailment rate exceeds 3%, the power command of the electrolyzer is immediately increased. The integral term eliminates long-term accumulated deviations; when the proportion of green hydrogen is insufficient for two consecutive hours, the proportion of hydrogen produced from renewable energy is gradually increased. The derivative term predicts the trend of deviation changes and adjusts the energy storage charging and discharging strategy in advance to smooth out power fluctuations. The adjusted coordination quantity is pushed to each agent in real time through a message queue, triggering a new round of local optimization.

[0066] In some embodiments, please refer to Figure 1 The optimization objective of the park model is: ; in, For electricity purchase price, For the power purchase capacity, For electricity sales price, For electricity sales capacity, This is the energy storage depreciation cost coefficient. For energy storage charging power, For energy storage discharge power, The penalty coefficient for wind and solar power curtailment. This refers to the power of wind and solar power that has been curtailed.

[0067] Total cost = Electricity purchase cost - Electricity sales revenue + Energy storage depreciation cost + Curtailment penalty cost.

[0068] The electricity purchase price represents the price paid for a unit of electricity from the grid at any given time, typically a time-of-use price, expressed in yuan / kWh. It directly impacts the system's electricity purchase decisions during periods of power shortage. During peak electricity price periods, the system will prioritize the use of energy storage or reduce unnecessary loads to lower electricity purchase costs.

[0069] Purchased power represents the real-time power purchased from the grid at any given time, measured in kW. It reflects the system's dependence on grid power; a higher value indicates insufficient renewable energy output or excessive load demand, necessitating the purchase of external power to meet needs.

[0070] The electricity sales price represents the price paid for a unit of electricity sold to the grid at any given time, expressed in yuan / kWh. The electricity sales price is typically lower than the electricity purchase price, but selling electricity helps recover excess renewable energy and reduces wind and solar curtailment. When there is excess photovoltaic output, selling electricity at a low price helps prevent overcharging of energy storage systems.

[0071] Electricity sold represents the real-time power sold to the grid at any given moment, measured in kW. This parameter reflects how the system handles excess electricity; a higher value indicates that renewable energy output exceeds local load and energy storage demand, requiring consumption through electricity sales.

[0072] The energy storage depreciation cost factor represents the depreciation cost of each unit of electricity charged and discharged by the energy storage device, expressed in yuan / kWh. Optimizing charging and discharging strategies can extend battery life.

[0073] Energy storage charging power represents the charging power of an energy storage device at any given time, measured in kW. A higher charging power means a stronger ability to absorb excess electricity, but it also increases depreciation costs. Priority charging is given when the risk of wind curtailment is high.

[0074] Energy storage discharge power represents the discharge power of an energy storage device at any given time, measured in kW. Higher discharge power means greater load support capability, but it also shortens battery life. Discharging during peak electricity price periods helps reduce electricity purchase costs.

[0075] The wind and solar curtailment penalty coefficient represents the penalty cost for each unit of renewable energy power curtailed (e.g., 1 kWh), expressed in yuan / kWh.

[0076] Curtailment power represents the amount of wind or solar power that cannot be absorbed at any given time, measured in kW. It reflects the utilization rate of renewable energy; a higher value indicates a larger load forecasting deviation in the dispatching strategy.

[0077] This invention proposes a system for green electricity to hydrogen production based on multi-agent source-grid-load-storage collaborative scheduling. Please refer to [link to relevant documentation]. Figure 2 ,include: Global unit 100 is configured to model the source, grid, load, and storage of the wind-solar-hydrogen-green power park to obtain the park model; Local unit 200 is configured to divide the park model into two time scales: day-ahead and intraday. It abstracts the equipment in the park model into several agents, builds a corresponding local optimization model for each agent, and couples all local optimization models through a preset coordination quantity to obtain a collaborative model. The scheduling scheme unit 300 is configured to acquire day-ahead information, input it into the collaborative model, obtain the day-ahead collaborative scheduling scheme, construct the intraday rolling optimization time window, and use the day-ahead collaborative scheduling scheme as a constraint to generate the intraday rolling scheduling scheme. The adaptive unit 400 is configured to execute an intraday rolling scheduling scheme to obtain the running results, evaluate the running results to obtain the running deviation, adaptively adjust the coordination amount based on the running deviation until the running deviation is lower than a preset value, and output the corresponding intraday rolling scheduling scheme.

[0078] In some embodiments, please refer to Figure 5 , Figure 6 , Figure 7 and Figure 8 One implementation of the present invention includes the following: Figure 8 The steps shown are as follows: Step S1: Modeling and parameter configuration of source-grid-load-storage in the park, such as... Figure 5 As shown; Model the distribution network topology of the wind-solar-hydrogen-green power park to determine the upstream grid interface points, transformers, feeders and key nodes.

[0079] A unified model is used to model photovoltaic power plants, wind farms, electrochemical energy storage devices, and water electrolysis hydrogen production devices, configuring rated capacity, upper and lower power limits, efficiency parameters, start-up and shutdown constraints, and operating cost parameters.

[0080] Configure the maximum power received and the maximum reverse power transmitted between the park and the upper-level power grid; configure the adjustable power range of hydrogen production load in different time periods, the allowable peak shifting time window, the cumulative hydrogen production demand and green hydrogen ratio requirements within the dispatch cycle; configure the upper and lower limits of energy storage SOC, initial and final SOC, maximum charge and discharge power and cycle depreciation cost.

[0081] Step S2: Multi-timescale partitioning Two time scales are set up: the day-ahead layer has a time resolution of 15 minutes and covers 96 time periods from 00:15 to 24:00 of the next day; the day-intraday layer has a time resolution of 15 minutes and performs rolling optimization every 15 minutes, with each roll covering the next 4 hours; and the boundary conditions and allowable deviation range between the day-ahead and day-intraday layers are determined.

[0082] Step S3: Multi-agent partitioning and local optimization model construction Different types of equipment within the park are abstracted into multiple autonomous agents, including: Source Agent_S: Represents photovoltaic power plants and wind farms; Agent_G: Represents the park's power distribution network and its interface with the upstream power grid; Hydrogen production load agent Agent_H: represents a water electrolysis hydrogen production unit or electrolyzer array; Energy Storage Agent_E: Represents an electrochemical energy storage system.

[0083] Build local optimization models for each agent: The source agent determines the renewable energy utilization power and the curtailed power under a given coordination amount, so as to reduce the penalty cost of wind and solar curtailment; The hydrogen production load agent determines the hydrogen production power for each time period under the conditions of meeting the hydrogen production task, peak shifting window and process constraints; Energy storage agents can smooth peak and valley loads and absorb excess wind and solar power within the SOC safe range through charging and discharging. The grid agent is responsible for node power balancing, line safety constraints, and power purchase and sale decisions.

[0084] The design coordinates the power decisions of each agent with the power balance constraints of the network side through the coordination, providing a foundation for distributed collaborative optimization.

[0085] Step S4: Multi-agent cooperative optimization scheduling Obtain the next day's photovoltaic and wind power output forecast curves and forecast ranges, and obtain the park's load, electricity price / carbon price, hydrogen production demand, and green hydrogen ratio targets.

[0086] The network agent generates initial day-ahead plans for electricity purchase and sale, energy storage SOC trajectory, and hydrogen production load based on forecast information.

[0087] Under the initial coordination parameters, the source agent, hydrogen production load agent, and energy storage agent independently solve their respective local optimization problems. The network agent aggregates the plans of each agent, checks the power balance and the degree of constraint violation, and updates the coordination parameters.

[0088] Repeated local optimization and coordination updates until convergence are obtained, resulting in a day-ahead coordinated dispatch scheme that includes power purchase and sale, energy storage SOC curve, hydrogen production load power curve, and wind and solar curtailment.

[0089] Step S5: Intraday Rolling Coordinated Scheduling The system acquires short-term forecasts for wind and solar power, electricity price changes, hydrogen production demand adjustments, and actual operating status based on a set rolling cycle.

[0090] Construct a rolling optimization time window covering the next 4 hours, using the day-ahead plan as a reference trajectory or soft constraint to limit the intraday adjustment range and constrain the state of SOC, hydrogen production, etc. at the end of the window to be close to the day-ahead target.

[0091] The multi-agent collaborative optimization is re-executed within the rolling time window to generate an intraday rolling scheduling scheme.

[0092] The scheduling instructions for the earliest segment of the time window are issued, and the above process is repeated as the time moves forward.

[0093] Step S6: Performance Deviation Assessment and Adaptive Parameter Update Statistical assessments were conducted on indicators such as electricity purchase and sale costs, wind and solar curtailment, green hydrogen ratio, electricity consumption per unit of hydrogen production, and carbon emissions.

[0094] Based on the deviation results, the target weights, constraint boundaries, and coordination algorithm parameters are adjusted to update the adjustable range of hydrogen production load and the safety boundary of energy storage SOC.

[0095] The scheduling results and parameter configurations under typical operating conditions are solidified into policy entries, and a policy library is built for subsequent scheduling initialization and policy recommendation.

[0096] Mathematical Modeling and Formulas Let the uniform time step be Δt, and the time set be T={1,2,…,T}.

[0097] The main decision variables include: Photovoltaic and wind power utilization capacity; Power purchased / sold; Energy storage charging and discharging power; Energy storage state of charge; Hydrogen production load power; Power curtailed from wind and solar power.

[0098] 1. Overall objective function (economic efficiency + curtailment) (1) in For electricity purchase / sale price, This is the energy storage depreciation cost coefficient. This represents the penalty coefficient for wind and solar power curtailment.

[0099] If carbon reduction and green hydrogen targets need to be explicitly considered, carbon emissions can be... ratio of green hydrogen Introducing the objective function: (2) in These are the weighting coefficients.

[0100] 2. Node power balance constraints (3) 3. Wind and solar power output and curtailment constraints (4) (5) 4. Energy Storage SOC Update and Constraints (6) (7) (8) in For the rated capacity of energy storage, These represent the charge / discharge efficiency, respectively.

[0101] 5. Hydrogen production load and hydrogen production constraints Hydrogen production per time period: (9) Hydrogen production task constraints for the day (or within the scheduling cycle): (10) Hydrogen production load power and peak shifting window constraints: (11) in The set of times during which operation / peak shifting is permitted.

[0102] 6. Power Constraints at the Grid Interface (12) 7. Green hydrogen ratio index Electricity generated from renewable energy sources in hydrogen production: (13) Total electricity consumption for hydrogen production: (14) The green hydrogen ratio is defined as: (15) Constraints can be set: (16) Multi-agent distributed cooperative optimization mechanism Treating the global problem as a collaborative optimization problem involving multiple agents, let the agent set be... These represent the source, grid, hydrogen production load, and energy storage agent, respectively. as an agent The set of decision variables, Let it be its local objective function. Let the equivalent power variable be at the node, then the global problem can be written as: (17) Construct the Lagrangian function: (18) in The Lagrange multiplier for nodal power balance constraints can be interpreted as the equivalent nodal electricity price or coordination quantity. Therefore, given... In this case, each agent can solve independently: (19) The network agent calculates and updates the power imbalance based on the power plans reported by each agent. Global coordination is achieved through iteration. This invention does not limit the specific iterative algorithm and can adopt methods such as Lagrange relaxation, ADMM, and gradient update.

[0103] At the day-ahead level, based on the next day's wind and solar forecasts, electricity price curves, and preliminary hydrogen production demand, the global optimization problem described by formulas (1) to (12) is constructed and solved through a multi-agent collaborative optimization mechanism to obtain the day-ahead reference power curves for photovoltaic, wind power, energy storage, hydrogen production, and the upper-level power grid.

[0104] In the intraday layer, the forecast data for the next 4 hours is updated every 15 minutes. Taking the day-ahead curve as a reference, a penalty term for deviation from the day-ahead plan is added to the objective function. Based on formulas (7) and (10), constraints are imposed on the window end state of SOC and cumulative hydrogen production. The updated intraday scheduling scheme is obtained by rolling solution.

[0105] After the operation cycle is completed, the green hydrogen ratio is calculated according to formulas (13) to (16). The statistics include electricity purchase and sale costs, wind and solar curtailment, and carbon emissions per unit of hydrogen production. If the economic efficiency, utilization rate, or green hydrogen ratio deviates significantly from expectations, the weights in formula (2) will be adjusted. , Together with the constraint parameters in formulas (11) and (16), an adaptive scheduling is formed.

[0106] System structure and Figure 6 , Figure 7 Corresponding, including: Data Acquisition and Management Module: Collects operational data from photovoltaic, wind power, hydrogen production, energy storage systems, power distribution networks, and upstream power grid interfaces within the park, and stores and preprocesses the data.

[0107] Prediction and Scenario Generation Module: Calls existing prediction models to generate wind and solar power output forecasts, park load / hydrogen production demand forecasts, and related scenarios at various time scales.

[0108] Multi-timescale scheduling module: includes two sub-modules, day-ahead and intraday, which perform collaborative scheduling according to steps S4 to S5 described above.

[0109] Multi-agent collaborative optimization module: including source agent module, network agent module, hydrogen production load agent module and energy storage agent module. Each module constructs local objectives according to formulas (17) and (18) and obtains a consistent scheduling scheme through coordination quantity iteration.

[0110] Execution and monitoring module: Converts scheduling results into control commands and sends them to field control equipment, and monitors the execution effect and operating status in real time.

[0111] Evaluation and Adaptive Optimization Module: Based on the operation results, evaluate the economic efficiency, renewable energy consumption rate, green hydrogen ratio and carbon emission indicators, and adaptively adjust the scheduling parameters in conjunction with the strategy library.

[0112] Based on the same inventive concept, according to another aspect of the present invention, such as Figure 3 As shown, an embodiment of the present invention also provides a computer device 30, which includes a processor 310 and a memory 320. The memory 320 stores a computer program 321 that can be run on the processor. When the processor 310 executes the program, it performs the steps of the method described above.

[0113] Based on the same inventive concept, according to another aspect of the present invention, such as Figure 4 As shown, embodiments of the present invention also provide a computer-readable storage medium 40, which stores a computer program 410 that, when executed by a processor, performs the methods described above.

[0114] Embodiments of the present invention may also include a corresponding computer device. The computer device includes a memory, at least one processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes any of the methods described above when executing the program.

[0115] The memory, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as program instructions / modules in the embodiments of this application. The processor executes various functional applications and data processing of the device by running the non-volatile software programs, instructions, and modules stored in the memory, thereby implementing the above-described method.

[0116] The memory may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the device. Furthermore, the memory may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In embodiments, the memory may optionally include memory remotely located relative to the processor, which can be connected to the local module via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0117] Finally, it should be noted that those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium for the program can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc. The above computer program embodiments can achieve the same or similar effects as any of the corresponding foregoing method embodiments.

[0118] Those skilled in the art will also understand that the various exemplary logic blocks, modules, circuits, and algorithm steps described in conjunction with the disclosure herein can be implemented as electronic hardware, computer software, or a combination of both. To clearly illustrate this interchangeability between hardware and software, the functionality of various illustrative components, blocks, modules, circuits, and steps has been generally described. Whether this functionality is implemented as software or as hardware depends on the specific application and the design constraints imposed on the system as a whole. Those skilled in the art can implement the functionality in various ways for each specific application, but such implementation decisions should not be construed as departing from the scope of the embodiments disclosed herein.

[0119] The above are exemplary embodiments disclosed in this invention. However, it should be noted that various changes and modifications can be made without departing from the scope of the embodiments of this invention as defined by the claims. The functions, steps, and / or actions of the methods according to the disclosed embodiments described herein do not need to be performed in any particular order. The sequence numbers of the disclosed embodiments of this invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments. Furthermore, although the elements disclosed in the embodiments of this invention may be described or claimed individually, they may be understood as multiple unless explicitly limited to a singular number.

[0120] It should be understood that, as used herein, the singular form "one" is intended to include the plural form as well, unless the context clearly supports an exception. It should also be understood that, as used herein, "and / or" refers to any and all possible combinations of one or more of the associated listed items.

[0121] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of the invention (including the claims) is limited to these examples. Within the framework of the invention, technical features of the above embodiments or different embodiments can be combined, and many other variations of different aspects of the invention exist, which are not provided in the details for the sake of brevity. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the invention should be included within the protection scope of the invention.

Claims

1. A method for green electricity to produce hydrogen based on multi-agent source-grid-load-storage collaborative scheduling, characterized in that, include: A model of the power generation, grid, load, and storage system in a wind-solar-hydrogen-to-green power industrial park is obtained. The park model is divided into two time scales: day-ahead and intraday. The equipment in the park model is abstracted into several agents. A corresponding local optimization model is built for each agent. The collaborative model is obtained by coupling all local optimization models through a preset coordination quantity. Obtain the day-ahead information and input it into the collaborative model to obtain the day-ahead collaborative scheduling scheme. Construct the intraday rolling optimization time window and use the day-ahead collaborative scheduling scheme as a constraint to generate the intraday rolling scheduling scheme. The intraday rolling scheduling scheme is executed to obtain the running results. The running results are evaluated to obtain the running deviation. The coordination amount is adaptively adjusted based on the running deviation until the running deviation is lower than the preset value, and the corresponding intraday rolling scheduling scheme is output.

2. The method for green electricity production based on multi-agent source-grid-load-storage coordinated scheduling according to claim 1, characterized in that, The steps for modeling the power generation, grid, load, and storage systems of the wind-solar-hydrogen-to-green power park to obtain the park model include: Construct a distribution network topology for the upstream grid interface points and key nodes in the wind-solar-hydrogen-green power industrial park; Modeling is performed using the distribution network topology as the network equipment, photovoltaic power plants and wind farms as the source equipment, electrochemical energy storage devices as the storage equipment, and water electrolysis for hydrogen production as the load equipment. The parameters between the wind-solar hydrogen production green power park and the upstream power grid, the hydrogen production load parameters of the water electrolysis hydrogen production device, and the energy storage parameters of the electrochemical energy storage device are configured to obtain the park model.

3. The method for green electricity production based on multi-agent source-grid-load-storage coordinated scheduling according to claim 2, characterized in that, The steps of abstracting and dividing the equipment in the park model into several agents, constructing a corresponding local optimization model for each agent, and coupling all local optimization models through a preset coordination quantity to obtain a collaborative model include: The equipment in the park model is abstracted and divided into several agents. Among them, the photovoltaic power station and wind farm are source agents, the park distribution network and its interface with the upper-level power grid are grid agents, the water electrolysis hydrogen production device is hydrogen production load agent, and the electrochemical energy storage device is energy storage agent. Build local optimization models for each agent; The optimization objective of the source agent is to calculate the decision between the utilization power of renewable energy and the curtailed power under a given coordination quantity; The optimization objective of the hydrogen production load proxy is to calculate the hydrogen production power decision for each time period; The optimization objective of the energy storage agent is to calculate the charging and discharging power decision within the SOC safe range; The optimization objective of the network agent is to calculate the power balance constraints of each agent; A collaborative model is obtained by coupling the power decisions of each agent with the power balance constraints of the network agent through a preset coordination quantity.

4. The method for green electricity production based on multi-agent source-grid-load-storage coordinated scheduling according to claim 3, characterized in that, The steps of obtaining day-ahead information and inputting it into the cooperative model to obtain the day-ahead cooperative scheduling scheme include: Obtain day-ahead information, which includes next-day photovoltaic and wind power output forecasts, park load, electricity / carbon price, hydrogen production demand, and green hydrogen ratio targets; Input the day-ahead information into the collaborative model, and the network agent generates an initial day-ahead plan based on the day-ahead information; Based on the preset coordination quantity, the source agent, hydrogen production load agent and energy storage agent in the collaborative model solve the local optimization problem respectively, and the network agent summarizes the results of each agent and updates the coordination quantity. Repeat the local optimization and update the coordination quantity until convergence, and use the day-ahead plan at the time of convergence as the day-ahead coordination scheduling scheme. The day-ahead coordinated scheduling scheme includes, but is not limited to, the power output curves of wind and solar power, the power purchased and sold, the SOC curve of energy storage, the power output curve of hydrogen production load, and the amount of wind and solar curtailment.

5. The method for green electricity production based on multi-agent source-grid-load-storage coordinated scheduling according to claim 4, characterized in that, The step of constructing the intraday rolling optimization time window and using the previous day's collaborative scheduling scheme as a constraint to generate the intraday rolling scheduling scheme includes: Construct a rolling time window for intraday optimization covering the next 4 hours; The aforementioned day-ahead coordinated scheduling scheme is used as a constraint; Within the intraday rolling optimization time window, the collaborative model is re-controlled for local optimization to generate an intraday rolling scheduling scheme.

6. The method for green electricity production based on multi-agent source-grid-load-storage coordinated scheduling according to claim 1, characterized in that, The steps of executing the intraday rolling scheduling scheme to obtain operational results, evaluating the operational results to obtain operational deviations, and adaptively adjusting the coordination quantity based on the operational deviations include: The results of executing the intraday rolling scheduling plan were obtained. The operational deviations are obtained by evaluating the operation results using indicators such as electricity purchase and sale costs, wind and solar curtailment, green hydrogen ratio, unit hydrogen production electricity consumption and carbon emissions. Based on operational deviations, the coordination quantity is adaptively adjusted. Record scheduling results and corresponding intraday rolling scheduling schemes, and build a strategy library.

7. The method for green electricity production based on multi-agent source-grid-load-storage coordinated scheduling according to claim 1, characterized in that, The optimization objective of the park model is: ； in, For electricity purchase price, For the power purchase capacity, For electricity sales price, For electricity sales capacity, This is the energy storage depreciation cost coefficient. For energy storage charging power, For energy storage discharge power, The penalty coefficient for wind and solar power curtailment. This refers to the power of wind and solar power that has been curtailed.

8. A system for green electricity-to-hydrogen production based on multi-agent source-grid-load-storage coordinated scheduling, characterized in that, include: The global unit is configured to model the source, grid, load, and storage of the wind-solar-hydrogen-green power park to obtain the park model; The local unit is configured to divide the park model into two time scales: day-ahead and intraday. The equipment in the park model is abstracted into several agents, and a corresponding local optimization model is built for each agent. A collaborative model is obtained by coupling all local optimization models through a preset coordination quantity. The scheduling scheme unit is configured to acquire day-ahead information, input it into the collaborative model, obtain the day-ahead collaborative scheduling scheme, construct the intraday rolling optimization time window, and use the day-ahead collaborative scheduling scheme as a constraint to generate the intraday rolling scheduling scheme. The adaptive unit is configured to execute an intraday rolling scheduling scheme to obtain the running results, evaluate the running results to obtain the running deviation, adaptively adjust the coordination amount based on the running deviation until the running deviation is lower than a preset value, and output the corresponding intraday rolling scheduling scheme.

9. A computer device, comprising: At least one processor; The processor also includes a memory storing a computer program that can run on the processor, characterized in that the processor executes the program to perform the steps of the method and system for green electricity production based on multi-agent source-grid-load-storage collaborative scheduling as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it performs the steps of the method and system for green electricity production based on multi-agent source-grid-load-storage collaborative scheduling as described in any one of claims 1 to 7.

Images