Multi-time-scale electro-hydrogen thermal coupling wind-solar hydrogen storage system scheduling and configuration method

Abstract

The invention relates to a multi-time-scale electro-hydrogen thermal coupling wind-light hydrogen storage system scheduling and configuration method, and belongs to the technical field of comprehensive energy system optimization. In order to solve the problems of time scale splitting, single energy and mass conversion, insufficient real-time response and the like of a traditional wind and light hydrogen storage system, a three-layer closed-loop collaborative framework of a cross-time scale optimization layer, an energy and mass coupling hub module and a cross-layer-energy and mass coordination module is constructed; and integrated optimization of annual / seasonal capacity configuration, weekly / monthly middle-term plans and intra-day / real-time scheduling is realized. Through explicit modeling electricity-hydrogen-heat multi-energy mass conversion and waste heat recycling, mixed integer programming and model prediction control technologies are combined, and the comprehensive energy utilization efficiency is remarkably improved while the operation stability of the system is guaranteed. The method adapts to the multi-time scale characteristics of wind and light output and load, can effectively reduce the operation cost of the system, reduces abandoned wind and light, and provides technical support for high-proportion renewable energy grid connection.

Description

Technical Field

[0001] This invention relates to the field of integrated energy system optimization scheduling and capacity planning technology, and in particular to a scheduling and configuration method for a multi-timescale electro-hydrogen-thermal coupled wind-solar-hydrogen storage system. Background Technology

[0002] As the penetration rate of renewable energy sources such as wind and solar continues to increase, the randomness and volatility of their output pose challenges to the stable operation of the power grid. Existing wind-solar + battery / hydrogen storage systems mostly adopt a two-layer optimization framework of "capacity configuration - intraday scheduling". The upper layer determines the installed capacity of equipment and energy storage through planning, while the lower layer carries out intraday economic scheduling within a given boundary. These methods usually only focus on the one-way electricity-hydrogen conversion link, with intraday economic optimization as the core objective.

[0003] However, there are significant seasonal / weekly structural differences between wind and solar resources and load demand. A single intraday optimization perspective cannot guarantee the matching of long-term capacity configuration with actual operation, easily leading to resource mismatch problems such as "over-powered" or equipment overload. Meanwhile, the electrolyzer hydrogen production and fuel cell power generation processes generate a large amount of recoverable waste heat, which existing models do not incorporate into energy flow management, resulting in low overall system energy utilization efficiency. Furthermore, traditional static scheduling mechanisms lack real-time rolling optimization capabilities, resulting in delayed responses to short-term sudden changes such as wind and solar power drops and load surges, easily leading to problems such as wind and solar curtailment, energy shortages, or soaring external purchase costs.

[0004] Currently, the industry urgently needs a comprehensive energy system optimization method that can take into account multi-timescale coordination, deep coupling of multiple energy types, and real-time fluctuation response, in order to overcome the bottlenecks of existing technologies in terms of long-term planning compatibility, energy utilization efficiency, and operational stability. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a scheduling and configuration method for a multi-timescale electro-hydrogen-thermal coupled wind-solar-hydrogen storage system. This method solves the problems of poor multi-timescale adaptability, low energy and quality utilization, insufficient real-time fluctuation response, and isolated cross-layer constraint coordination in traditional wind-solar-hydrogen storage systems. By constructing a three-layer closed-loop collaborative framework consisting of a cross-timescale optimization layer, an energy and quality coupling hub module, and a cross-layer-energy and quality coordination module, it achieves integrated optimization of annual / quarterly capacity configuration, weekly / monthly medium-term planning, and daily / real-time scheduling. This realizes integrated closed-loop optimization of capacity configuration, medium-term planning, and real-time scheduling, as well as deep coupling of multiple energy qualities (electricity, hydrogen, and heat). Through explicit modeling of electro-hydrogen-thermal multi-energy quality conversion and waste heat recovery, combined with mixed integer programming and model predictive control techniques, it significantly improves overall energy utilization efficiency while ensuring system operational stability. The technical solution of this invention is highly reproducible, adaptable to the multi-timescale characteristics of wind and solar power output and load, effectively reduces system operating costs, reduces wind and solar curtailment, and provides technical support for high-proportion renewable energy grid connection.

[0006] To address the aforementioned technical problems, this invention provides the following technical solution: a method for scheduling and configuring a multi-timescale electro-hydrogen-thermal coupled wind-solar-hydrogen storage system, comprising the following steps: S1. Construct an electro-hydrogen-thermal coupled wind-solar-hydrogen storage system and obtain system structural parameters and boundaries, including wind power / solar installed capacity range, battery / hydrogen storage / thermal storage capacity range, rated power and efficiency of electrolyzer / fuel cell, and recovery coefficient; at the same time, obtain representative period wind and solar power output, load and heat load, electricity price / hydrogen price data and discretize them into multi-time scale datasets. S2. In the upper-level capacity configuration module, using capacity vectors A multi-objective optimization model at the annual / quarterly scale was constructed for the decision variables, and the non-dominated sorting genetic algorithm NSGA-II was used to search for a Pareto capacity solution set that satisfies the constraints. S3. In the mid-level medium-term scheduling module, based on capacity vectors and medium-to-long-term weekly / monthly scenario data, using planning parameters... A weekly / monthly scale planning optimization model was constructed for the decision variables, and the optimized planning parameters were obtained by solving the Particle Swarm Optimization (PSO) algorithm. Output contracts / budgets, reference inventory levels, and reserve margins; S4. In the lower-level short-term real-time scheduling module, based on the capacity vector, optimized planning parameters, short-term forecast data, and current system state, a model predictive control (MPC) rolling mechanism is adopted: at each rolling moment... Construct a prediction window The mixed-integer linear programming (MILP) scheduling model is used and solved by CPLEX to obtain the control sequence. Execute the first control variable And update the system status; S5. Construct a cross-layer-energy quality coordination module. Based on the key performance indicators (KPIs) output by the lower layer, the constraint default statistics, or the solution of the shadow price of the optimal solution of the dual variable, the module is used to dynamically correct the planning parameters, update the penalty factor in the upper layer capacity evaluation function, or return to the upper layer to re-evaluate and screen the candidate capacity solutions, forming a closed-loop collaborative optimization.

[0007] Furthermore, in the cross-layer-energy quality coordination module, the key performance indicators (KPIs) include power curtailment rate, heat load shortage rate, number of equipment start-ups and shutdowns, and energy purchase cost deviation; the constraint default statistics include the number of times battery limits are exceeded, the number of times hydrogen storage inventory limits are exceeded, and the number of times thermal storage inventory limits are exceeded.

[0008] Furthermore, the specific steps of the upper-layer capacity configuration module include: With capacity vector Let be the decision variables, where For wind power installed capacity, For photovoltaic installed capacity, For the rated power of the electrolytic cell, For fuel cell rated power, For battery energy storage capacity, For hydrogen storage capacity, Thermal energy storage capacity; The initial population of individuals is randomly generated, and the capacity vector of each individual satisfies the upper and lower bound constraints. With the goal of minimizing annualized total cost, minimizing wind and solar curtailment rate, and minimizing external energy purchase cost, the non-dominated sorting genetic algorithm NSGA-II is used to perform non-dominated sorting, crowding calculation, selection / crossover / mutation operations on the population. After the iteration, the set of individuals with the first non-dominated level is the Pareto capacity solution set. Each candidate capacity solution needs to call the mid-level medium-term scheduling module and the lower-level short-term real-time scheduling module to calculate its operational key performance indicators (KPIs). If the curtailment rate or heat load shortage rate of a candidate capacity solution exceeds a preset threshold, it is removed from the Pareto capacity solution set. If the number of equipment start-ups and shutdowns of a candidate capacity solution exceeds a preset threshold, the middle and lower level evaluators are called to recalculate its KPIs and re-sort it according to the updated penalty factor. If multiple candidate solutions have constraint violations, the upper and lower bound constraints of capacity configuration are adjusted, the candidate solutions are re-evaluated, and a new Pareto capacity solution set is output.

[0009] Furthermore, the specific steps of the intermediate-level scheduling module include: Obtain the capacity vector, medium- and long-term weekly / monthly scenario data, and operation completion data output from the upper-level capacity configuration. The medium- and long-term weekly / monthly scenario data includes wind and solar power output forecasts, electricity / heat load forecasts, and energy price forecasts. The operation completion data includes the upper and lower limits of inventory and the range of reserve margin. The planned parameters are represented using low-dimensional parameterization, including the medium-term power / hydrogen purchase contract volume, battery SOC, reference trajectory control points for hydrogen storage and thermal storage inventory, and reserve margin; wherein, the reference trajectory control points for battery SOC, hydrogen storage inventory, and thermal storage inventory are set based on the selection of key time points to set their target states, and the obtained reference trajectory control points are used to generate a complete time series through linear interpolation or smoothing functions. The objective function is to minimize the medium- and long-term operating costs, and the low-dimensional parameters of the planning parameters are used as particle positions to search for the optimal planning parameters through PSO iteration. The optimal planning parameters obtained through optimization are projected and repaired to ensure that they are within the capacity configuration boundary and operational constraints.

[0010] Furthermore, the specific steps of the lower-level short-term real-time scheduling module include: At each scrolling moment The system obtains current system status and short-term forecast data. The current system status includes battery SOC, hydrogen storage inventory, and thermal storage inventory. The short-term forecast data includes wind and solar power output forecast, electricity / heat load forecast, and energy price forecast. Constructing a prediction window With control window ; Electrolytic cell power Fuel cell power Power purchased Electricity sales capacity Battery charging power Battery discharge power Hydrogen storage and charging flow rate Hydrogen storage and hydrogen release flow rates Thermal energy storage power Thermal energy storage heat release power As a continuous variable; the mutually exclusive state of electricity purchase and sale; Battery charging and discharging mutual exclusion state Electrolytic cell start-up and shutdown status Fuel cell start-stop status As a binary variable; The MILP constraints are constructed using three energy quality balance constraints, energy storage dynamic update constraints, equipment mutual exclusion constraints, equipment start-up and shutdown constraints, ramp-up constraints, and upper and lower bound constraints of inventory and power. Among them, the three energy quality balance constraints include electric power balance constraints, hydrogen balance constraints, and thermal balance constraints; the energy storage dynamic update constraints include battery SOC update, hydrogen storage inventory update, and thermal storage inventory update; the equipment mutual exclusion constraints include mutual exclusion of electricity purchase and sale and mutual exclusion of battery charge and discharge; and the equipment start-up and shutdown constraints include binding the start-up and shutdown of the electrolyzer / fuel cell to the power boundary. The MILP model is solved using CPLEX branching strategies employing the strongest branch or pseudo-cost branch, and pruning conditions including bounded pruning when the lower bound of node relaxation is no better than the current optimal upper bound, and feasible pruning when relaxation is infeasible. The resulting model outputs the control sequence within the prediction window. Execute the first control variable It also updates the battery SOC and the hydrogen / thermal storage system status, and uses the updated status as the initial status for the next rolling time k+1; After executing the short-term real-time scheduling control, the KPI, constraint default statistics, and shadow price are output to the cross-layer-energy quality coordination module. Based on the feedback information, the cross-layer-energy quality coordination module determines whether to trigger adjustments to the planning parameters, update the penalty factor, or re-select candidate solutions to adjust the middle and upper layers. If the planning parameter is modified, the middle layer updates the planning parameters and sends them to the lower layer. If the penalty factor is updated, the upper layer adjusts the capacity evaluation function and re-evaluates the candidate capacity solutions. If re-selection is triggered, the upper layer removes infeasible solutions and outputs a new Pareto solution set. The lower layer re-executes the scheduling based on the new planning parameters or candidate capacity solutions, outputting new feedback information to form a closed-loop optimization.

[0011] Furthermore, the specific formulas for the three-energy mass balance constraint and the energy storage dynamic update constraint in the MILP constraint construction are as follows: Electric power balance constraints: ; in, This represents the wind power output at time t. Represents the photovoltaic power at time t. This represents the power of wind and solar power curtailment at time t. This represents the purchased power at time t. This represents the electricity sold at time t. This represents the system's electrical load demand at time t; Hydrogen balance constraints: ; in, This represents the amount of hydrogen produced by the electrolyzer at time t. This represents the amount of hydrogen the system purchased from external sources at time t. This represents the amount of hydrogen the system sells to the outside at time t. represents the amount of hydrogen consumed by the fuel cell at time t, and represents the system hydrogen load demand at time t; Thermal equilibrium constraint: ; in, This represents the amount of waste heat recovered by the electrolytic cell at time t. This represents the amount of waste heat recovered from the fuel cell at time t. This represents the system heat load demand at time t; Battery SOC Update: ; in, This indicates the state of charge of the battery energy storage unit at time t. This indicates the energy conversion efficiency during the charging process. This indicates the energy conversion efficiency during the discharge process; Hydrogen storage inventory update: ; in, This indicates the inventory level of the hydrogen storage unit at time t; Thermal storage inventory update: ; in, This indicates the inventory level of the thermal energy storage tank at time t. These refer to the heat conversion efficiency during heat storage and heat release processes, respectively. For time step.

[0012] Furthermore, the specific formulas for the device mutual exclusion constraints, device start / stop constraints, and ramp constraints in the MILP constraint construction are as follows: Electricity purchase and sale are mutually exclusive: , ;in, Let t be the purchased power. Let t be the power sold at time t. For a sufficiently large constant, For a mutually exclusive state of electricity purchase and sale, the binary variables are: 1 = electricity purchase, 0 = electricity sale; Battery charge and discharge mutual exclusion: ; ; in, Maximum charging power for the battery. This is the battery's maximum discharge power. The battery charging and discharging mutually exclusive states are represented by binary variables: 1 = charging, 0 = discharging. Electrolytic cell start-stop state variables Binding constraints to power boundaries: ; in , These are the minimum and maximum operating power of the electrolytic cell, respectively. Let t be the power consumption of the electrolytic cell. The binary variable represents the start-up and stop status of the electrolytic cell: 1 = start-up, 0 = stop. Fuel cell start-stop state variables Binding constraints to power boundaries: ,in , These represent the minimum and maximum operating power of the fuel cell, respectively. Let t be the power consumption of the fuel cell. The binary variable represents the start-stop state of the fuel cell: 1 = start-up, 0 = stop. Climbing constraints: ; ; in, This represents the maximum downward slope rate of the electrolytic cell. This represents the maximum upward slope rate of the electrolytic cell. This represents the maximum downhill ramp rate for the fuel cell. This represents the maximum uphill ramp rate of the fuel cell.

[0013] Furthermore, the electro-hydrogen-thermal coupled wind-solar-hydrogen storage system includes at least a wind power generation unit, a photovoltaic power generation unit, a grid interface unit, a battery energy storage unit, an electrolyzer hydrogen production unit, a hydrogen storage unit, a fuel cell power generation unit, a waste heat recovery unit, a thermal energy storage unit, an electrical load unit, and a thermal load unit. Each unit achieves deep coupling of the three energy substances of electricity, hydrogen, and heat through an energy-mass coupling hub module. The energy-mass coupling hub module includes an electrolyzer hydrogen production model, an electrolyzer waste heat recovery model, a fuel cell power generation model, and a fuel cell waste heat recovery model.

[0014] Furthermore, the core model formula in the energy-mass coupling hub module is: Electrolyzer hydrogen production model: ;in, This represents the amount of hydrogen produced by the electrolyzer at time t. This indicates the efficiency of the electrolyzer in converting electrical energy into hydrogen energy. This represents the electrolytic cell power at time t. Represents the time interval for discretization. This indicates the lower heating value of hydrogen. Electrolytic cell waste heat recovery model: ,in The waste heat recovery coefficient of the electrolytic cell; Indicates the proportion of waste heat from the electrolytic cell; Fuel cell power generation model: ;in, This represents the fuel cell power at time t. This indicates the hydrogen conversion efficiency of the fuel cell. This represents the amount of hydrogen consumed by the fuel cell at time t; Fuel cell waste heat recovery model: ,in The waste heat recovery coefficient of the fuel cell. This indicates the proportion of waste heat from the fuel cell.

[0015] Furthermore, the specific priority rules for the dynamically adjusted plan parameters are as follows: When the heat load shortage rate is greater than or equal to the preset threshold, the charging priority of thermal energy storage is increased. When the shadow price of thermal energy storage inventory constraint obtained by CPLEX solution is greater than or equal to the preset threshold, dynamic adjustment of thermal energy storage reference inventory trajectory is triggered to prevent thermal load shortage in advance. When the curtailment rate is greater than or equal to the preset threshold, increase the upper limit of the reference inventory trajectory for energy storage / hydrogen storage; When the deviation in energy purchase cost is greater than or equal to a preset threshold, the volume of medium-term power / hydrogen purchase contracts will be adjusted.

[0016] By employing the above technical solutions, this invention provides a scheduling and configuration method for a multi-timescale electro-hydrogen-thermal coupled wind-solar-hydrogen storage system. Addressing the pain points of traditional wind-solar-hydrogen storage systems, such as fragmentation across multiple timescales, single energy-mass conversion, and insufficient real-time response, it constructs a closed-loop collaborative framework encompassing cross-timescale optimization, energy-mass coupling hubs, and cross-layer coordination. This achieves deep integration of capacity configuration, medium-term planning, and real-time scheduling, demonstrating significant advantages over existing technologies. 1. By optimizing the three-tiered collaborative approach of annual / quarterly, weekly / monthly, and daily / real-time, the time scale barrier of the traditional two-tiered framework is broken, enabling a deep match between long-term capacity planning and short-term operation scheduling. This reduces the annualized total cost of the system and the curtailment of wind and solar power, and solves the resource mismatch problem caused by prioritizing short-term over long-term. 2. By modeling the entire energy flow path from electricity to hydrogen, hydrogen to electricity, and waste heat recovery, the waste heat recovery rate of electrolyzers and fuel cells is improved, breaking through the overall energy utilization efficiency of the system. Waste heat recovery and thermal energy storage are explicitly introduced to improve overall energy efficiency and enhance the adaptability of heating scenarios, meeting the diversified load demands of electricity, heat, and hydrogen. 3. By adopting the MPC+MILP rolling scheduling mechanism and combining it with linear modeling of engineering constraints, it can respond to sudden changes in wind and solar power output and load demand within minutes, control the heat load shortage rate, reduce the number of equipment start-ups and shutdowns, improve the robustness of real-time fluctuations, significantly enhance the system's ability to cope with extreme operating conditions, and reduce the risk of power curtailment and shortage. 4. By referencing inventory trajectory and KPI feedback mechanisms, the conflict between medium-term plans and intraday strategies can be mitigated, thereby improving scheduling stability. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a schematic diagram of the structure of the three-energy-mass coupling wind-solar-hydrogen-storage-thermal complementary system of the present invention; Figure 2 This is a flowchart of the three-layer cross-timescale energy-mass coupling collaborative optimization interaction process of the present invention. Detailed Implementation

[0018] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. This will allow for a full understanding of how the present application uses technical means to solve technical problems and achieve technical effects, and to facilitate its implementation.

[0019] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0020] Please refer to Figures 1-2 This illustration demonstrates a specific implementation of this embodiment. This embodiment constructs a three-layer collaborative framework: annual / quarterly capacity configuration, weekly / monthly mid-term planning, and daily / real-time rolling scheduling. Dynamic feedback at each time scale is achieved through a cross-layer energy quality coordination module. Equipment parameters output from the upper-layer capacity configuration directly guide the mid-layer planning, and the mid-layer planning parameters provide a reference trajectory for the lower-layer real-time scheduling, avoiding the traditional problem of "disconnect between planning and operation" and achieving consistency between long-term planning and short-term operation. KPI indicators, constraint default statistics, and dual variables generated by the lower-layer real-time scheduling are fed back to the upper layer in real time, dynamically adjusting the penalty factor of capacity configuration and the mid-layer planning parameters, forming a full-cycle optimization closed loop of "planning-operation-correction-replanning." Annual-scale planning ensures the long-term economic efficiency of the system, weekly-scale planning balances mid-term resource allocation, and real-time scheduling responds to short-term fluctuations, achieving complementary advantages across different time scales.

[0021] The electro-hydrogen-thermal coupled wind-solar-hydrogen storage system of the present invention includes at least a wind power generation unit, a photovoltaic power generation unit, a grid interface unit, a battery energy storage unit, an electrolyzer hydrogen production unit, a hydrogen storage unit, a fuel cell power generation unit, a waste heat recovery unit, a thermal energy storage unit, an electrical load unit, and a thermal load unit. Each unit achieves deep coupling of the three energy substances (electricity, hydrogen, and heat) through an energy-mass coupling hub module. The energy-mass coupling hub module includes an electrolyzer hydrogen production model, an electrolyzer waste heat recovery model, a fuel cell power generation model, and a fuel cell waste heat recovery model. Wherein, as... Figure 1As shown, the wind power generation unit is connected to the grid interface unit, converting wind energy into AC power to provide clean energy input to the system; the photovoltaic power generation unit is connected to the grid interface unit, converting solar energy into DC power, which is then converted back into AC power by an inverter and injected into the system; the grid interface unit, as the system's energy hub, connects all energy input / output units, realizing the distribution, conversion, and scheduling of energy, and supporting the purchase and sale of electricity with the external grid (receiving energy input from wind power, photovoltaics, and fuel cells, distributing energy to electrolyzers, battery storage, and electrical loads, realizing bidirectional power exchange with the external grid); battery storage... The unit connects to the grid interface unit to smooth out fluctuations in renewable energy output, provide short-term power support, charge and store energy when wind / solar power output is excessive, and discharge to supply power during peak load periods or when renewable energy output is insufficient. It also participates in grid frequency regulation and peak shaving ancillary services. The electrolyzer hydrogen production unit obtains electrical energy from the grid interface unit, produces hydrogen which is then connected to the hydrogen storage unit. Waste heat is connected to the waste heat recovery unit, converting surplus electrical energy into hydrogen energy for long-term storage. This achieves the spatial and temporal transfer of electrical energy, absorbs abandoned wind and solar power, improves the utilization rate of renewable energy, produces hydrogen to meet the needs of fuel cell power generation and hydrogen load, and recovers hydrogen. Waste heat generated during electrolysis is used for heating; the hydrogen storage unit receives hydrogen from the electrolyzer's hydrogen production unit to provide hydrogen for the fuel cell power generation unit and hydrogen load unit, achieving large-scale, long-term hydrogen storage; the fuel cell power generation unit obtains hydrogen from the hydrogen storage unit, generates electricity which is connected to the grid interface unit, and waste heat is connected to the waste heat recovery unit to convert hydrogen energy into electrical energy, providing a stable power supply to the system when needed. It also provides electricity when renewable energy output is insufficient, responding to grid peak-shaving needs and providing flexible power support. Waste heat generated during power generation is recovered for heating; the waste heat recovery unit receives hydrogen from the electrolyzer... The waste heat from the fuel cell is used to generate heat energy which is then connected to the thermal energy storage unit or directly supplied to the heat load unit, recovering waste heat during the energy conversion process and improving the overall energy utilization efficiency of the system. The thermal energy storage unit obtains heat energy from the waste heat recovery unit and supplies heat energy to the heat load unit, realizing the storage and scheduling of heat energy and balancing fluctuations in heat load demand. The electrical load unit obtains electrical energy from the grid interface unit and consumes electrical energy to meet various electricity needs of users. The heat load unit obtains heat energy from the waste heat recovery unit, fuel cell power generation unit, electrolyzer hydrogen production unit, and thermal energy storage unit and consumes heat energy to meet users' heating, hot water, and other needs.

[0022] like Figure 2 As shown, the cross-layer-energy coordination module enables bidirectional communication between the control unit and all energy conversion and storage units, realizing global optimization scheduling and closed-loop control of the system. It receives the operating status data and KPI indicators of each unit, dynamically adjusts the operating parameters and power commands of each unit, realizes the optimal complementary utilization of the three energy forms of electricity, hydrogen and heat, and constructs a collaborative optimization mechanism of upper-layer capacity configuration, middle-layer medium-term scheduling and lower-layer short-term real-time scheduling. like Figure 2As shown, the energy-mass coupling hub module is the core for achieving deep coupling of electricity, hydrogen, and heat. It includes an electrolyzer hydrogen production model for converting electricity to hydrogen, an electrolyzer waste heat recovery model for recovering waste heat during the electrolysis process, a fuel cell power generation model for converting hydrogen to electricity, and a fuel cell waste heat recovery model for recovering waste heat during the fuel cell power generation process. Through these models, the system can explicitly model the entire chain of conversion paths from electricity to hydrogen (electrolyzer), hydrogen to electricity (fuel cell), electricity to heat (electrolyzer waste heat), and hydrogen to heat (fuel cell waste heat) based on real-time operating status and market signals. This allows for flexible adjustment of energy conversion, breaking through the limitations of traditional electricity-hydrogen conversion alone, and achieving cascaded energy utilization and optimal economic operation. Secondly, the waste heat generated by the electrolyzer and fuel cell is recovered through the waste heat recovery unit, stored in the thermal energy storage unit, and supplied to the heat load on demand, improving the overall energy utilization efficiency of the system. Simultaneously, it meets the diversified needs of electrical load, thermal load, and even hydrogen load, adapting to various application scenarios such as industrial parks, commercial complexes, and residential communities.

[0023] Specifically, the core model formula in the energy-mass coupling hub module is as follows: Electrolyzer hydrogen production model: ;in, This represents the amount of hydrogen produced by the electrolyzer at time t. This indicates the efficiency of the electrolyzer in converting electrical energy into hydrogen energy. This represents the electrolytic cell power at time t. Represents the time interval for discretization. This indicates the lower heating value of hydrogen. Electrolytic cell waste heat recovery model: ,in The waste heat recovery coefficient of the electrolytic cell; Indicates the proportion of waste heat from the electrolytic cell; Fuel cell power generation model: ;in, This represents the fuel cell power at time t. This indicates the hydrogen conversion efficiency of the fuel cell. This represents the amount of hydrogen consumed by the fuel cell at time t; Fuel cell waste heat recovery model: ,in The waste heat recovery coefficient of the fuel cell. This indicates the proportion of waste heat from the fuel cell.

[0024] Please refer to Figure 1 This embodiment proposes a scheduling and configuration method for a multi-timescale electro-hydrogen-thermal coupled wind-solar-hydrogen storage system, which includes the following steps: S1. Construct an electro-hydrogen-thermal coupled wind-solar-hydrogen storage system and obtain system structural parameters and boundaries, including wind / solar installed capacity range, battery / hydrogen storage / thermal storage capacity range, rated power and efficiency of electrolyzer / fuel cell, and recovery coefficient; simultaneously obtain representative period wind and solar power output, load and heat load, electricity / hydrogen price data and discretize them into multi-timescale datasets; specifically, the selection principle for representative period data is: covering the seasonal characteristics of typical years, the lower layer uses Δt=15min (or 5min) discretization; the middle layer uses weekly / monthly periods to form representative days or weeks; the upper layer uses annual / quarterly periods as representative periods, and data sources include historical operating data, public databases or prediction model outputs.

[0025] S2. In the upper-level capacity configuration module, using capacity vectors A multi-objective optimization model at the annual / quarterly scale was constructed for the decision variables, and the non-dominated sorting genetic algorithm NSGA-II was used to search for a Pareto capacity solution set that satisfies the constraints. As a preferred embodiment of step S2, the specific steps of the upper-layer capacity configuration module include: With capacity vector For decision variables, capacity vector Real number encoding is used, where For wind power installed capacity, For photovoltaic installed capacity, For the rated power of the electrolytic cell, For fuel cell rated power, For battery energy storage capacity, For hydrogen storage capacity, Thermal energy storage capacity; Randomly generated An initial population of individuals, each with a capacity vector that satisfies upper and lower bound constraints; With multiple optimization objectives such as minimizing annualized total cost, minimizing wind and solar curtailment rate, minimizing external energy purchase cost, or minimizing carbon emissions, the non-dominated sorting genetic algorithm NSGA-II is used to perform non-dominated sorting, crowding calculation, selection / crossover / mutation operations on the population, with a number of iterations. The crossover probability Probability of mutation After the iteration is complete, the set of individuals at the first non-dominated level is the Pareto capacity solution set. Each candidate capacity solution needs to call the mid-level medium-term scheduling module and the lower-level short-term real-time scheduling module to calculate its operational key performance indicators (KPIs). If the curtailment rate or heat load shortage rate of a candidate capacity solution exceeds a preset threshold, it is removed from the Pareto capacity solution set. If the number of equipment start-ups and shutdowns of a candidate capacity solution exceeds a preset threshold, the middle and lower level evaluators are invoked to recalculate its KPIs and re-rank it according to the updated penalty factor. If multiple candidate solutions have constraint violations, the upper and lower bound constraints of capacity configuration are adjusted, the candidate solutions are re-evaluated, and a new Pareto capacity solution set is output. Constraint handling is as follows: feasible solutions are given priority, and infeasible solutions are penalized according to the degree of violation. Fitness evaluation is performed on each candidate solution... Get by calling the middle-level PSO Then, the lower-level MPC+CPLEX is called to return the annualized running KPI to calculate the objective function.

[0026] S3. In the mid-level medium-term scheduling module, based on capacity vectors and medium-to-long-term weekly / monthly scenario data, using planning parameters... A weekly / monthly scale planning optimization model was constructed for the decision variables, and the optimized planning parameters were obtained by solving the Particle Swarm Optimization (PSO) algorithm. Output contracts / budgets, reference inventory levels, and reserve margins; As a preferred embodiment of step S3, the specific steps of the mid-level intermediate scheduling module include: Obtain the capacity vector, medium- and long-term weekly / monthly scenario data, and operation completion data output from the upper-level capacity configuration. The medium- and long-term weekly / monthly scenario data includes wind and solar power output forecasts, electricity / heat load forecasts, and energy price forecasts. The operation completion data includes the upper and lower limits of inventory and the range of reserve margin. The planned parameters are represented using low-dimensional parameterization, including the medium-term electricity / hydrogen purchase contract volume, battery SOC, reference trajectory control points for hydrogen storage and thermal storage inventory, and reserve margin; wherein, the reference trajectory control points for battery SOC, hydrogen storage inventory, and thermal storage inventory are set based on selecting key time points to set their target states, for example, giving SOC / S on a daily basis. h2 / S th The reference trajectory control points are obtained and linearly interpolated to the intraday level; the monthly power / hydrogen purchase contract volume and reserve margin are also given; the complete time series is generated from the obtained reference trajectory control points through linear interpolation or smoothing function; The Particle Swarm Optimization (PSO) algorithm was used as the optimization tool, with settings for the number of particles and inertia weights. Learning factors Number of iterations The objective function is to minimize the medium- and long-term operating costs, and the low-dimensional parameters of the planning parameters are used as particle positions to search for the optimal planning parameters through PSO iteration. The optimal planning parameters obtained through optimization are projected and repaired to ensure that they are within the capacity configuration boundary and operational constraints.

[0027] S4. In the lower-level short-term real-time scheduling module, based on the capacity vector, optimized planning parameters, short-term forecast data, and current system state, a model predictive control (MPC) rolling mechanism is adopted: at each rolling moment... Construct a prediction window The mixed-integer linear programming (MILP) scheduling model is used and solved by CPLEX to obtain the control sequence. Execute the first control variable And update the system status; As a preferred embodiment of step S4, the specific steps of the lower-level short-term real-time scheduling module include: At each scrolling moment The system obtains current system status and short-term forecast data. The current system status includes battery SOC, hydrogen storage inventory, and thermal storage inventory. The short-term forecast data includes wind and solar power output forecast, electricity / heat load forecast, and energy price forecast. Constructing a prediction window With control window ,in (Corresponding to 4-12 hours); Electrolytic cell power Fuel cell power Power purchased Electricity sales capacity Battery charging power Battery discharge power Hydrogen storage and charging flow rate Hydrogen storage and hydrogen release flow rates Thermal energy storage power Thermal energy storage heat release power As a continuous variable; the mutually exclusive state of electricity purchase and sale; Battery charging and discharging mutual exclusion state Electrolytic cell start-up and shutdown status Fuel cell start-stop status As a binary variable; The MILP constraints are constructed using three energy quality balance constraints, energy storage dynamic update constraints, equipment mutual exclusion constraints, equipment start-up and shutdown constraints, ramp-up constraints, and upper and lower bound constraints of inventory and power. Among them, the three energy quality balance constraints include electric power balance constraints, hydrogen balance constraints, and thermal balance constraints; the energy storage dynamic update constraints include battery SOC update, hydrogen storage inventory update, and thermal storage inventory update; the equipment mutual exclusion constraints include mutual exclusion of electricity purchase and sale and mutual exclusion of battery charge and discharge; and the equipment start-up and shutdown constraints include binding the start-up and shutdown of the electrolyzer / fuel cell to the power boundary. Specifically, power balance constraints: ; in, This represents the wind power output at time t. Represents the photovoltaic power at time t. This represents the power of wind and solar power curtailment at time t. This represents the purchased power at time t. This represents the electricity sold at time t. This represents the system's electrical load demand at time t; Hydrogen balance constraints: ; in, This represents the amount of hydrogen produced by the electrolyzer at time t. This represents the amount of hydrogen the system purchased from external sources at time t. This represents the amount of hydrogen the system sells to the outside at time t. represents the amount of hydrogen consumed by the fuel cell at time t, and represents the system hydrogen load demand at time t; Thermal equilibrium constraint: ; in, This represents the amount of waste heat recovered by the electrolytic cell at time t. This represents the amount of waste heat recovered from the fuel cell at time t. This represents the system heat load demand at time t; Battery SOC Update: ; in, This indicates the state of charge of the battery energy storage unit at time t. This indicates the energy conversion efficiency during the charging process. This indicates the energy conversion efficiency during the discharge process; Hydrogen storage inventory update: ; in, This indicates the inventory level of the hydrogen storage unit at time t; Thermal storage inventory update: ; in, This indicates the inventory level of the thermal energy storage tank at time t. These refer to the heat conversion efficiency during heat storage and heat release processes, respectively. For time step.

[0028] Electricity purchase and sale are mutually exclusive: , ;in, Let t be the purchased power. Let t be the power sold at time t. For a sufficiently large constant, For a mutually exclusive state of electricity purchase and sale, the binary variables are: 1 = electricity purchase, 0 = electricity sale; Battery charge and discharge mutual exclusion: ; ; in, Maximum charging power for the battery. This is the battery's maximum discharge power. The battery charging and discharging mutually exclusive states are represented by binary variables: 1 = charging, 0 = discharging. Electrolytic cell start-stop state variables Binding constraints to power boundaries: ; in , These are the minimum and maximum operating power of the electrolytic cell, respectively. Let t be the power consumption of the electrolytic cell. The binary variable represents the start-up and stop status of the electrolytic cell: 1 = start-up, 0 = stop. Fuel cell start-stop state variables Binding constraints to power boundaries: ,in , These represent the minimum and maximum operating power of the fuel cell, respectively. Let t be the power consumption of the fuel cell. The binary variable represents the start-stop state of the fuel cell: 1 = start-up, 0 = stop. Climbing constraints: ; ; in, This represents the maximum downward slope rate of the electrolytic cell. This represents the maximum upward slope rate of the electrolytic cell. This represents the maximum downhill ramp rate for the fuel cell. This represents the maximum uphill ramp rate of the fuel cell.

[0029] Soft constraints can be applied to the deviation from the reference inventory trajectory, and penalties can be added to the objective function. The objective function may include energy purchase cost, power curtailment penalty, start-up and shutdown penalty, and deviation penalty from the reference trajectory. The MILP model is solved using CPLEX with the strongest branch or pseudo-cost branch strategy, and pruning conditions such as boundary pruning when the lower bound of node relaxation is not better than the current optimal upper bound, and feasibility pruning when relaxation is not feasible. A time limit of TimeLimit and MIPGap are set. The solution stops when MIPGap ≤ threshold ε or when the time limit of TimeLimit is reached, ensuring that each rolling cycle can be solved within the specified time. The control sequence within the prediction window is output. Execute the first control variable The battery SOC and hydrogen / thermal storage system status are updated, and the updated status is used as the initial status for the next rolling time k+1; where MIPGap = 1e-4~1e-3; TimeLimit is set according to real-time requirements; the solution priority is set according to feasibility priority / optimal priority to ensure the stability of rolling solution; After executing the short-term real-time scheduling control, the KPI, constraint default statistics, and shadow price are output to the cross-layer-energy quality coordination module. Based on the feedback information, the cross-layer-energy quality coordination module determines whether to trigger adjustments to the planning parameters, update the penalty factor, or re-select candidate solutions to adjust the middle and upper layers. If the planning parameter is modified, the middle layer updates the planning parameters and sends them to the lower layer. If the penalty factor is updated, the upper layer adjusts the capacity evaluation function and re-evaluates the candidate capacity solutions. If re-selection is triggered, the upper layer removes infeasible solutions and outputs a new Pareto solution set. The lower layer re-executes the scheduling based on the new planning parameters or candidate capacity solutions, outputting new feedback information to form a closed-loop optimization.

[0030] Specifically, the priority rules for the dynamic adjustment plan parameters are as follows: When the heat load shortage rate is greater than or equal to the preset threshold, the charging priority of thermal energy storage is increased. When the shadow price of thermal energy storage inventory constraint obtained by CPLEX solution is greater than or equal to the preset threshold, dynamic adjustment of thermal energy storage reference inventory trajectory is triggered to prevent thermal load shortage in advance. When the curtailment rate is greater than or equal to the preset threshold, increase the upper limit of the reference inventory trajectory for energy storage / hydrogen storage; When the deviation in energy purchase cost is greater than or equal to a preset threshold, the volume of medium-term power / hydrogen purchase contracts will be adjusted.

[0031] It should be noted that the preset thresholds include: power curtailment rate ≥8%, heat load shortage rate ≥5%, energy purchase cost deviation ≥10%, and equipment start-up and shutdown frequency ≥20 times per month. The thresholds can be set according to the actual situation.

[0032] In this embodiment, the invention employs a rolling scheduling mechanism combining Model Predictive Control (MPC) and Mixed Integer Linear Programming (MILP): MILP models are constructed and solved at each rolling time point to respond in real-time to sudden changes in wind and solar power output, load demand, and energy prices, thereby improving rapid response capabilities; engineering constraints such as mutual exclusion of power purchase and sale, mutual exclusion of battery charging and discharging, equipment start-up and shutdown, and power ramp-up are linearized to ensure the solvability of the model and the engineering feasibility of the solution; the strongest branch or pseudo-cost branch strategy of the CPLEX solver is adopted, combined with boundary pruning and feasibility pruning conditions, to shorten the solution time to the minute level while ensuring solution accuracy; through multi-period optimization within the prediction window, operational risks such as power curtailment, heat load shortage, and frequent equipment start-up and shutdown are predicted and avoided in advance, reducing the system's power curtailment rate and heat load shortage rate.

[0033] S5. Construct a cross-layer-energy quality coordination module. Based on the key performance indicators (KPIs) output by the lower layer, the constraint default statistics, or the solution of the shadow price of the optimal solution of the dual variable, the module is used to dynamically correct the planning parameters, update the penalty factor in the upper layer capacity evaluation function, or return to the upper layer to re-evaluate and screen the candidate capacity solutions, forming a closed-loop collaborative optimization.

[0034] As a preferred implementation of step S5, in the cross-layer-energy quality coordination module, the key performance indicators (KPIs) include curtailment rate, heat load shortage rate, number of equipment start-ups and shutdowns, and deviation in energy purchase costs; the constraint default statistics include the number of times batteries exceed limits, the number of times hydrogen storage inventory exceeds limits, and the number of times thermal storage inventory exceeds limits.

[0035] In this embodiment, the present invention achieves unified coordination of multiple time scales and multiple energy flows through a cross-layer-energy coordination module: using reference trajectories as guidance, the reference trajectories of battery SOC, hydrogen storage inventory, and thermal storage inventory output by the mid-layer plan provide operating boundaries for the lower-layer real-time scheduling, avoiding the randomness of short-term scheduling; using KPI indicators such as curtailment rate, thermal load shortage rate, and equipment start-up and shutdown frequency generated by the lower-layer real-time scheduling as feedback drive, the mid-layer planning parameters and upper-layer capacity configuration penalty factors are dynamically adjusted; using the constraint dual variables (shadow prices) obtained by CPLEX solution to identify system bottlenecks, prioritizing the supply of critical loads, and improving system operational stability; by dynamically correcting planning parameters and capacity evaluation functions, the strategy conflict between mid-term planning and intraday real-time scheduling is effectively alleviated, the system operational stability is greatly improved, and cross-layer conflicts are mitigated.

[0036] This invention outputs an upper-level Pareto capacity solution set and recommended capacity, a middle-level planning parameter θ (contract quantity, reference inventory trajectory, reserve margin), and a lower-level real-time control sequence u(t) and operational KPIs (cost, curtailment, start-up / shutdown frequency, supply shortage statistics, etc.). Through three-level collaborative optimization at the annual / quarterly, weekly / monthly, and daily / real-time levels, it breaks the time scale barrier of the traditional two-level framework, enabling deep matching between long-term capacity planning and short-term operation scheduling, reducing the system's annualized total cost and curtailment of wind and solar power, and solving the resource mismatch problem caused by prioritizing short-term over long-term. By modeling the entire energy flow path from electricity to hydrogen, hydrogen to electricity, and waste heat recovery, it integrates the electrolyzer and fuel cell... The improved waste heat recovery rate of the pool breaks through the overall energy utilization efficiency of the system. The explicit introduction of waste heat recovery and thermal energy storage enhances the overall energy efficiency and improves the adaptability of heating scenarios, meeting the diversified load demands of electricity, heat, and hydrogen. The adoption of the MPC+MILP rolling scheduling mechanism, combined with linearized modeling of engineering constraints, can respond to sudden changes in wind and solar output and load demand within minutes. The heat load shortage rate is controlled, the number of equipment start-ups and shutdowns is reduced, and the robustness to real-time fluctuations is improved, significantly enhancing the system's ability to cope with extreme operating conditions and reducing the risk of power curtailment and shortages. By referencing inventory trajectories and KPI feedback mechanisms, the conflict between medium-term plans and intraday strategies is mitigated, improving scheduling stability.

[0037] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0038] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus or device (such as a computer-based system, a processor-included system or other system that can fetch and execute instructions from, an instruction execution system, apparatus or device).

[0039] The above embodiments provide a detailed description of the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for scheduling and configuring a multi-timescale electro-hydrogen-thermal coupled wind-solar-hydrogen storage system, characterized in that, Includes the following steps: S1. Construct an electro-hydrogen-thermal coupled wind-solar-hydrogen storage system and obtain system structural parameters and boundaries, including wind power / solar installed capacity range, battery / hydrogen storage / thermal storage capacity range, rated power and efficiency of electrolyzer / fuel cell, and recovery coefficient; at the same time, obtain representative period wind and solar power output, load and heat load, electricity price / hydrogen price data and discretize them into multi-time scale datasets. S2. In the upper-level capacity configuration module, using capacity vectors A multi-objective optimization model at the annual / quarterly scale was constructed for the decision variables, and the non-dominated sorting genetic algorithm NSGA-II was used to search for a Pareto capacity solution set that satisfies the constraints. S3. In the mid-level medium-term scheduling module, based on capacity vectors and medium-to-long-term weekly / monthly scenario data, using planning parameters... A weekly / monthly scale planning optimization model was constructed for the decision variables, and the optimized planning parameters were obtained by solving the Particle Swarm Optimization (PSO) algorithm. Output contracts / budgets, reference inventory levels, and reserve margins; S4. In the lower-level short-term real-time scheduling module, based on the capacity vector, optimized planning parameters, short-term forecast data, and current system state, a model predictive control (MPC) rolling mechanism is adopted: at each rolling moment... Construct a prediction window The mixed-integer linear programming (MILP) scheduling model is used and solved by CPLEX to obtain the control sequence. Execute the first control variable And update the system status; S5. Construct a cross-layer-energy quality coordination module. Based on the key performance indicators (KPIs) output by the lower layer, the constraint default statistics, or the solution of the shadow price of the optimal solution of the dual variable, the module is used to dynamically correct the planning parameters, update the penalty factor in the upper layer capacity evaluation function, or return to the upper layer to re-evaluate and screen the candidate capacity solutions, forming a closed-loop collaborative optimization.

2. The scheduling and configuration method for a multi-timescale electro-hydrogen-thermal coupled wind-solar-hydrogen storage system according to claim 1, characterized in that: In the cross-layer-energy quality coordination module, the key performance indicators (KPIs) include curtailment rate, heat load shortage rate, number of equipment start-ups and shutdowns, and deviation in energy purchase costs; the constraint default statistics include the number of times battery limits are exceeded, the number of times hydrogen storage inventory limits are exceeded, and the number of times thermal storage inventory limits are exceeded.

3. The scheduling and configuration method for a multi-timescale electro-hydrogen-thermal coupled wind-solar-hydrogen storage system according to claim 2, characterized in that: The specific steps of the upper-layer capacity configuration module include: With capacity vector Let be the decision variables, where For wind power installed capacity, For photovoltaic installed capacity, For the rated power of the electrolytic cell, For fuel cell rated power, For battery energy storage capacity, For hydrogen storage capacity, Thermal energy storage capacity; The initial population of individuals is randomly generated, and the capacity vector of each individual satisfies the upper and lower bound constraints. With the goal of minimizing annualized total cost, minimizing wind and solar curtailment rate, and minimizing external energy purchase cost, the non-dominated sorting genetic algorithm NSGA-II is used to perform non-dominated sorting, crowding calculation, selection / crossover / mutation operations on the population. After the iteration, the set of individuals with the first non-dominated level is the Pareto capacity solution set. Each candidate capacity solution needs to call the mid-level medium-term scheduling module and the lower-level short-term real-time scheduling module to calculate its operational key performance indicators (KPIs). If the curtailment rate or heat load shortage rate of a candidate capacity solution exceeds a preset threshold, it is removed from the Pareto capacity solution set. If the number of equipment start-ups and shutdowns of a candidate capacity solution exceeds a preset threshold, the middle and lower level evaluators are called to recalculate its KPIs and re-sort it according to the updated penalty factor. If multiple candidate solutions have constraint violations, the upper and lower bound constraints of capacity configuration are adjusted, the candidate solutions are re-evaluated, and a new Pareto capacity solution set is output.

4. The scheduling and configuration method for a multi-timescale electro-hydrogen-thermal coupled wind-solar-hydrogen storage system according to claim 3, characterized in that: The specific steps of the intermediate-level scheduling module include: Obtain the capacity vector, medium- and long-term weekly / monthly scenario data, and operation completion data output from the upper-level capacity configuration. The medium- and long-term weekly / monthly scenario data includes wind and solar power output forecasts, electricity / heat load forecasts, and energy price forecasts. The operation completion data includes the upper and lower limits of inventory and the range of reserve margin. The planned parameters are represented using low-dimensional parameterization, including the medium-term power / hydrogen purchase contract volume, battery SOC, reference trajectory control points for hydrogen storage and thermal storage inventory, and reserve margin; wherein, the reference trajectory control points for battery SOC, hydrogen storage inventory, and thermal storage inventory are set based on the selection of key time points to set their target states, and the obtained reference trajectory control points are used to generate a complete time series through linear interpolation or smoothing functions. The objective function is to minimize the medium- and long-term operating costs, and the low-dimensional parameters of the planning parameters are used as particle positions to search for the optimal planning parameters through PSO iteration. The optimal planning parameters obtained through optimization are projected and repaired to ensure that they are within the capacity configuration boundary and operational constraints.

5. The scheduling and configuration method for a multi-timescale electro-hydrogen-thermal coupled wind-solar-hydrogen storage system according to claim 1, characterized in that: The specific steps of the lower-level short-term real-time scheduling module include: At each scrolling moment The system obtains current system status and short-term forecast data. The current system status includes battery SOC, hydrogen storage inventory, and thermal storage inventory. The short-term forecast data includes wind and solar power output forecast, electricity / heat load forecast, and energy price forecast. Constructing a prediction window With control window ; Electrolytic cell power Fuel cell power Power purchased Electricity sales capacity Battery charging power Battery discharge power Hydrogen storage and charging flow rate Hydrogen storage and hydrogen release flow rates Thermal energy storage power Thermal energy storage heat release power As a continuous variable; the mutually exclusive state of electricity purchase and sale; Battery charging and discharging mutual exclusion state Electrolytic cell start-up and shutdown status Fuel cell start-stop status As a binary variable; The MILP constraints are constructed using three energy quality balance constraints, energy storage dynamic update constraints, equipment mutual exclusion constraints, equipment start-up and shutdown constraints, ramp-up constraints, and upper and lower bound constraints of inventory and power. Among them, the three energy quality balance constraints include electric power balance constraints, hydrogen balance constraints, and thermal balance constraints; the energy storage dynamic update constraints include battery SOC update, hydrogen storage inventory update, and thermal storage inventory update; the equipment mutual exclusion constraints include mutual exclusion of electricity purchase and sale and mutual exclusion of battery charge and discharge; and the equipment start-up and shutdown constraints include binding the start-up and shutdown of the electrolyzer / fuel cell to the power boundary. The MILP model is solved using CPLEX branching strategies employing the strongest branch or pseudo-cost branch, and pruning conditions including bounded pruning when the lower bound of node relaxation is no better than the current optimal upper bound, and feasible pruning when relaxation is infeasible. The resulting model outputs the control sequence within the prediction window. Execute the first control variable It also updates the battery SOC and the hydrogen / thermal storage system status, and uses the updated status as the initial status for the next rolling time k+1; After executing the short-term real-time scheduling control, the KPI, constraint default statistics, and shadow price are output to the cross-layer-energy quality coordination module. Based on the feedback information, the cross-layer-energy quality coordination module determines whether to trigger adjustments to the planning parameters, update the penalty factor, or re-select candidate solutions to adjust the middle and upper layers. If the planning parameter is modified, the middle layer updates the planning parameters and sends them to the lower layer. If the penalty factor is updated, the upper layer adjusts the capacity evaluation function and re-evaluates the candidate capacity solutions. If re-selection is triggered, the upper layer removes infeasible solutions and outputs a new Pareto solution set. The lower layer re-executes the scheduling based on the new planning parameters or candidate capacity solutions, outputting new feedback information to form a closed-loop optimization.

6. The scheduling and configuration method for a multi-timescale electro-hydrogen-thermal coupled wind-solar-hydrogen storage system according to claim 5, characterized in that: The specific formulas for the three-energy mass balance constraint and the energy storage dynamic update constraint in the MILP constraint construction are as follows: Electric power balance constraints: ; in, This represents the wind power output at time t. Represents the photovoltaic power at time t. This represents the power of wind and solar power curtailment at time t. This represents the purchased power at time t. This represents the electricity sold at time t. This represents the system's electrical load demand at time t; Hydrogen balance constraints: ; in, This represents the amount of hydrogen produced by the electrolyzer at time t. This represents the amount of hydrogen the system purchased from external sources at time t. This represents the amount of hydrogen the system sells to the outside at time t. represents the amount of hydrogen consumed by the fuel cell at time t, and represents the system hydrogen load demand at time t; Thermal equilibrium constraint: ; in, This represents the amount of waste heat recovered by the electrolytic cell at time t. This represents the amount of waste heat recovered from the fuel cell at time t. This represents the system heat load demand at time t; Battery SOC update: ; in, This indicates the state of charge of the battery energy storage unit at time t. This indicates the energy conversion efficiency during the charging process. This indicates the energy conversion efficiency during the discharge process; Hydrogen storage inventory update: ; in, This indicates the inventory level of the hydrogen storage unit at time t; Thermal storage inventory update: ; in, This indicates the inventory level of the thermal energy storage tank at time t. These refer to the heat conversion efficiency during heat storage and heat release processes, respectively. For time step.

7. The scheduling and configuration method for a multi-timescale electro-hydrogen-thermal coupled wind-solar-hydrogen storage system according to claim 5, characterized in that: The specific formulas for the device mutual exclusion constraints, device start / stop constraints, and ramp constraints in the MILP constraint construction are as follows: Electricity purchase and sale are mutually exclusive: , ;in, Let t be the purchased power. Let t be the power sold at time t. For a sufficiently large constant, For a mutually exclusive state of electricity purchase and sale, the binary variables are: 1 = electricity purchase, 0 = electricity sale; Battery charge and discharge mutual exclusion: ; ； in, Maximum charging power for the battery. This is the battery's maximum discharge power. The battery charging and discharging mutually exclusive states are represented by binary variables: 1 = charging, 0 = discharging. Electrolytic cell start-stop state variables Binding constraints to power boundaries: ; in , These are the minimum and maximum operating power of the electrolytic cell, respectively. Let t be the power consumption of the electrolytic cell. The binary variable represents the start-up and stop status of the electrolytic cell: 1 = start-up, 0 = stop. Fuel cell start-stop state variables Binding constraints to power boundaries: ,in , These represent the minimum and maximum operating power of the fuel cell, respectively. Let t be the power consumption of the fuel cell. The binary variable represents the start-stop state of the fuel cell: 1 = start-up, 0 = stop. Climbing constraints: ; ； in, This represents the maximum downward slope rate of the electrolytic cell. This represents the maximum upward slope rate of the electrolytic cell. This represents the maximum downhill ramp rate for the fuel cell. This represents the maximum uphill ramp rate of the fuel cell.

8. The scheduling and configuration method for a multi-timescale electro-hydrogen-thermal coupled wind-solar-hydrogen storage system according to claim 1, characterized in that: The electro-hydrogen-thermal coupled wind-solar-hydrogen storage system includes at least a wind power generation unit, a photovoltaic power generation unit, a grid interface unit, a battery energy storage unit, an electrolyzer hydrogen production unit, a hydrogen storage unit, a fuel cell power generation unit, a waste heat recovery unit, a thermal energy storage unit, an electrical load unit, and a thermal load unit. Each unit achieves deep coupling of the three energy substances of electricity, hydrogen, and heat through an energy-mass coupling hub module. The energy-mass coupling hub module includes an electrolyzer hydrogen production model, an electrolyzer waste heat recovery model, a fuel cell power generation model, and a fuel cell waste heat recovery model.

9. The scheduling and configuration method for a multi-timescale electro-hydrogen-thermal coupled wind-solar-hydrogen storage system according to claim 8, characterized in that: The core model formula in the energy-mass coupling hub module is: Electrolyzer hydrogen production model: ;in, This represents the amount of hydrogen produced by the electrolyzer at time t. This indicates the efficiency of the electrolyzer in converting electrical energy into hydrogen energy. This represents the electrolytic cell power at time t. Represents the time interval for discretization. This indicates the lower heating value of hydrogen. Electrolytic cell waste heat recovery model: ,in The waste heat recovery coefficient of the electrolytic cell; Indicates the proportion of waste heat from the electrolytic cell; Fuel cell power generation model: ;in, This represents the fuel cell power at time t. This indicates the hydrogen conversion efficiency of the fuel cell. This represents the amount of hydrogen consumed by the fuel cell at time t; Fuel cell waste heat recovery model: ,in The waste heat recovery coefficient of the fuel cell. This indicates the proportion of waste heat from the fuel cell.

10. A scheduling and configuration method for a multi-timescale electro-hydrogen-thermal coupled wind-solar-hydrogen storage system according to any one of claims 1-9, characterized in that: The specific priority rules for the dynamic adjustment plan parameters are as follows: When the heat load shortage rate is greater than or equal to the preset threshold, the charging priority of thermal energy storage is increased. When the shadow price of thermal energy storage inventory constraint obtained by CPLEX solution is greater than or equal to the preset threshold, dynamic adjustment of thermal energy storage reference inventory trajectory is triggered to prevent thermal load shortage in advance. When the curtailment rate is greater than or equal to the preset threshold, increase the upper limit of the reference inventory trajectory for energy storage / hydrogen storage; When the deviation in energy purchase cost is greater than or equal to a preset threshold, the volume of medium-term power / hydrogen purchase contracts will be adjusted.

Images