A method and device for optimizing the configuration of the double-layer capacity of a large-scale wind-solar-storage hydrogen production system

Abstract

The embodiment of the application provides a large-scale wind-solar-storage hydrogen production double-layer capacity configuration optimization method and device, a double-layer capacity configuration optimization model is constructed: an upper-layer capacity optimization model is used for solving an optimal capacity configuration scheme of each energy equipment, and a lower-layer scheduling optimization model is used for solving an optimal scheduling scheme under a given capacity configuration, wherein comprehensive fluctuation is composed of multiple factors which are weighted and reflect electrolytic cell start-stop characteristics and working condition conversion stability. A first optimization algorithm (non-dominated sorting genetic algorithm) is used for solving the upper-layer multi-objective problem, and a non-dominated solution set composed of multiple capacity configuration schemes is obtained; for each configuration scheme in the solution set, a second optimization algorithm (particle swarm optimization algorithm) is used for solving the lower-layer scheduling problem, and an optimal scheduling scheme and operation cost corresponding to the configuration scheme are obtained. Finally, a multi-attribute decision method is used for comprehensively evaluating all capacity configuration schemes and scheduling schemes, and a target capacity configuration scheme and a corresponding scheduling strategy are selected from the capacity configuration schemes.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of clean energy and hydrogen energy utilization technology, and in particular to a method and apparatus for optimizing the capacity configuration of a large-scale wind-solar-storage-hydrogen production dual-layer system. Background Technology

[0002] Large-scale production of green hydrogen from renewable energy sources such as wind and solar power is a key pathway to achieving a low-carbon transition in the energy system. However, the intermittency and volatility of wind and solar power output directly lead to drastic fluctuations in hydrogen production power, severely impacting the operating efficiency, lifespan, and hydrogen production stability of electrolyzers. To address this issue, existing technologies have attempted to quantify the power fluctuations of electrolyzers to optimize system operation. For example, statistical indicators such as standard deviation and coefficient of variation are used to describe the magnitude of power fluctuations, or a penalty cost mechanism is introduced to indirectly characterize the impact of fluctuations on economics.

[0003] However, existing technologies still have significant shortcomings and deficiencies in the following aspects: First, existing methods for quantifying electrolyzer fluctuations are mostly limited to a single dimension (such as fluctuation amplitude), failing to comprehensively characterize the coupled operational characteristics and overall fluctuation features of multiple electrolyzers under complex operating conditions (such as cold start, hot start, standby, and fluctuating operation). In particular, there is a lack of a comprehensive quantitative index that can reflect the start-up and shutdown interference, operating condition transition stability, and fluctuation continuity of the electrolyzer array, resulting in an inaccurate characterization of the dynamic operating state of the hydrogen production system. Second, some existing technologies use a penalty cost mechanism to indirectly quantify the impact of fluctuations. The setting of the penalty coefficient in this mechanism often relies on human experience, which is highly subjective and difficult to objectively and accurately match the nonlinear and time-varying impact of fluctuations on equipment lifespan and system efficiency under actual operating conditions, thus failing to provide accurate economic guidance for scheduling decisions. Third, traditional single-layer capacity configuration optimization frameworks usually only focus on economic optimization at the long-term planning level, failing to effectively couple long-term capacity planning with short-term operation scheduling. This prevents the system from fully leveraging the synergistic potential of wind and solar energy complementarity and multi-energy regulation of electricity and hydrogen storage when dealing with rapid intraday fluctuations in wind and solar resources and dynamic changes in hydrogen load. Consequently, the optimized results result in low wind and solar energy absorption rates and poor hydrogen supply stability in actual operation, making it difficult to achieve the overall economic goals. Summary of the Invention

[0004] This invention provides a method and apparatus for optimizing the dual-layer capacity configuration of wind, solar, energy storage and hydrogen production on a large scale. This addresses the technical problem that existing large-scale wind, solar, energy storage and hydrogen production systems cannot simultaneously consider the objective quantification of the multi-dimensional fluctuation characteristics of the electrolyzer array, the utilization of the complementary characteristics of the electro-hydrogen coupling of multiple subsystems, and the coordinated optimization of long-term capacity planning and short-term operation scheduling. This results in the system being unable to achieve a comprehensive optimal balance between economy, reliability and wind and solar energy absorption rate.

[0005] In a first aspect, embodiments of the present invention provide a method for optimizing the capacity configuration of a large-scale wind-solar-storage-hydrogen production dual-layer system, comprising: S1. Based on pre-acquired wind and solar power output scenarios and hydrogen load scenarios, construct a large-scale wind-solar-storage-hydrogen production system model. This model includes a power subsystem, a hydrogen subsystem, and an electrolyzer array. S2. Construct a two-layer capacity configuration optimization model, comprising an upper-layer capacity optimization model and a lower-layer scheduling optimization model. S3. The upper-layer capacity optimization model aims to minimize annualized total cost, minimize curtailment rate, and maximize energy supply reliability, and sets equipment capacity configuration constraints to solve for the capacity configuration schemes of each energy device in the power subsystem, hydrogen subsystem, and electrolyzer array. S4. The lower-layer scheduling optimization model aims to minimize the overall volatility of the electrolyzer array. The system sets operational constraints to solve for the optimal scheduling scheme under a given capacity configuration scheme; wherein, the comprehensive volatility index is obtained by weighted calculation of multiple factors characterizing the start-up and shutdown characteristics and operating condition transition stability of the electrolytic cell array; S5, the upper-level capacity optimization model is solved using the first optimization algorithm to obtain a non-dominated solution set including multiple capacity configuration schemes; for each capacity configuration scheme in the non-dominated solution set, the lower-level scheduling optimization model is solved using the second optimization algorithm to obtain the corresponding optimal scheduling scheme and operating cost; S6, based on all capacity configuration schemes and the corresponding optimal scheduling schemes, a target capacity configuration scheme and the corresponding optimal scheduling scheme are selected from the non-dominated solution set using a multi-attribute decision method.

[0006] Secondly, embodiments of the present invention provide a large-scale wind-solar-storage-hydrogen production dual-layer capacity configuration optimization device, comprising: The scenario construction module is used to construct a large-scale wind-solar-storage-hydrogen production system model based on pre-acquired wind and solar power output scenarios and hydrogen load scenarios. The large-scale wind-solar-storage-hydrogen production system model includes a power subsystem, a hydrogen subsystem, and an electrolyzer array. A two-layer model construction module is used to construct a two-layer capacity configuration optimization model, which includes an upper-layer capacity optimization model and a lower-layer scheduling optimization model. The upper-level optimization module is used to optimize the annualized total cost, the curtailment rate, and the energy supply reliability, and sets equipment capacity configuration constraints to solve the capacity configuration scheme of each energy device in the power subsystem, hydrogen subsystem, and electrolyzer array. The lower-level optimization module is used to minimize the overall volatility of the electrolytic cell array as the optimization objective and to set system operation constraints to solve the optimal scheduling scheme under a given capacity configuration scheme; wherein, the overall volatility index is obtained by weighted calculation of multiple factors characterizing the start-up and shutdown characteristics and the stability of operating condition transitions of the electrolytic cell array. The two-layer solution module is used to solve the upper-layer capacity optimization model using a first optimization algorithm to obtain a non-dominated solution set including multiple capacity configuration schemes. For each capacity configuration scheme in the non-dominated solution set, the second optimization algorithm is used to solve the lower-layer scheduling optimization model to obtain the corresponding optimal scheduling scheme and operating cost. The decision output module is used to select the target capacity configuration scheme and the corresponding optimal scheduling scheme from the non-dominated solution set using a multi-attribute decision method based on all capacity configuration schemes and the corresponding optimal scheduling schemes.

[0007] This invention provides a method and apparatus for optimizing the dual-layer capacity configuration of wind, solar, energy storage, and hydrogen production. A dual-layer capacity configuration optimization model is constructed: the upper-layer model aims to minimize annualized total cost, curtailment rate, and power supply reliability, and sets equipment capacity constraints to solve for the optimal capacity configuration scheme of each energy device; the lower-layer scheduling optimization model aims to minimize the overall volatility of the electrolyzer array, and sets system operation constraints to solve for the optimal scheduling scheme under a given capacity configuration. The overall volatility is composed of a weighted average of multiple factors reflecting the start-up and shutdown characteristics and operational condition transition stability of the electrolyzers. A first optimization algorithm (non-dominated sorting genetic algorithm) is used to solve the upper-layer multi-objective problem, obtaining a non-dominated solution set consisting of multiple capacity configuration schemes. For each configuration scheme in the solution set, a second optimization algorithm (particle swarm optimization algorithm) is used to solve the lower-layer scheduling problem, obtaining the corresponding preferred scheduling scheme and operating cost. Finally, a multi-attribute decision-making method is used to comprehensively evaluate all capacity configuration schemes and their scheduling schemes, selecting the target capacity configuration scheme and its corresponding scheduling strategy. Attached Figure Description

[0008] 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 some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a flowchart illustrating the method for optimizing the capacity configuration of a large-scale wind-solar-storage-hydrogen production dual-layer system provided in an embodiment of the present invention. Figure 2 A logic diagram for decomposition and modeling of the fluctuation characteristics of an electrolytic cell array provided in an embodiment of the present invention; Figure 3 A flowchart for constructing the comprehensive volatility index of an electrolyzer based on the MIC-improved CRITIC method provided in this embodiment of the invention; Figure 4The flowchart of the solution of the double-layer nested optimization algorithm for the wind-solar-storage coupled hydrogen production system provided in the embodiments of the present invention is shown below. Figure 5 This is a structural block diagram of a large-scale wind-solar-storage-hydrogen production dual-layer capacity configuration optimization device provided in an embodiment of the present invention. Detailed Implementation

[0010] Figure 1 This is a flowchart of a method for optimizing the capacity configuration of a large-scale wind-solar-storage-hydrogen production dual-layer system according to an embodiment of the present invention, with reference to... Figure 1 The method includes: S1. Based on the pre-acquired wind and solar power output scenarios and hydrogen load scenarios, construct a large-scale wind, solar, storage and hydrogen production system model. The large-scale wind, solar, storage and hydrogen production system model includes a power subsystem, a hydrogen subsystem and an electrolyzer array.

[0011] Among them, the wind and solar power output scenario is an energy output scenario corresponding to typical wind and solar meteorological parameters obtained by processing historical wind and solar data collected on-site. The hydrogen load scenario is a typical energy load scenario obtained by processing historical hydrogen consumption data collected on-site. The large-scale wind, solar, storage and hydrogen production system model is a comprehensive energy system model that integrates the complementary characteristics of the power subsystem, hydrogen subsystem and electrolyzer array's electric-hydrogen coupling. The power subsystem is the core unit for realizing electricity production, storage and transmission, the hydrogen subsystem is the core unit for realizing hydrogen storage, and the electrolyzer array is the core unit for realizing the conversion of electricity into hydrogen energy. The core device for conversion; S1 in this embodiment is based on the pre-acquired wind and solar power output scenario and hydrogen load scenario, and combines the coupling and complementary characteristics of the power subsystem and the hydrogen subsystem to build a large-scale wind, solar, energy storage and hydrogen production system model including the power subsystem, the hydrogen subsystem and the electrolyzer array. In view of the shortcomings of the existing technology in not fully exploring the complementary potential of wind, solar and energy storage and hydrogen storage and the lack of electro-hydrogen coupling correlation in system modeling, it provides a precise system operation carrier for subsequent two-layer optimization, consolidates the foundation for the coordinated operation of the wind, solar and energy storage and hydrogen production system, and ensures that the system optimization model matches the actual operating conditions.

[0012] S2. Construct a two-layer capacity configuration optimization model, which includes an upper-layer capacity optimization model and a lower-layer scheduling optimization model.

[0013] The dual-layer capacity configuration optimization model is a hierarchical optimization framework that decouples long-term capacity planning from short-term operation scheduling. The upper-layer capacity optimization model is an optimization model that performs multi-objective planning for the installed capacity of system equipment, while the lower-layer scheduling optimization model is a scheduling model that optimizes the real-time operation strategy of the system. In this embodiment, S2 constructs a dual-layer capacity configuration optimization model that includes the upper-layer capacity optimization model and the lower-layer scheduling optimization model. To address the shortcomings of traditional single-layer optimization frameworks that only consider long-term planning and cannot adapt to short-term dynamic operating condition changes, the hierarchical architecture decouples the functions of capacity configuration and operation scheduling, establishes a bidirectional feedback mechanism between the two models, solves the problem of disconnect between long-term planning and short-term scheduling, and provides an optimization framework to support the balance between system economy and operational stability.

[0014] S3. The upper-level capacity optimization model takes the minimum annualized total cost, the minimum curtailment rate, and the maximum energy supply reliability as optimization objectives, and sets equipment capacity configuration constraints to solve the capacity configuration scheme of each energy device in the power subsystem, hydrogen subsystem, and electrolyzer array.

[0015] Among them, annualized total cost is the core indicator for measuring the economic efficiency of the system throughout its entire life cycle, including system investment cost, maintenance cost, and operating cost. Curtailment rate is an indicator for measuring the utilization efficiency of wind and solar energy, referring to the proportion of unutilized electricity generated by photovoltaic and wind power to the total power generation. Power supply reliability is an indicator for measuring the stability of the system's power supply, referring to the system's ability to meet electricity and hydrogen loads. Equipment capacity configuration constraints are the limiting conditions for the range of installed capacity, curtailment rate, power supply reliability, and maximum investment cost. The capacity configuration scheme is the combination of the installed capacity of each energy device in the power subsystem, hydrogen subsystem, and electrolyzer array. In this embodiment, S3 takes the minimum annualized total cost, minimum curtailment rate, and maximum power supply reliability of the upper-level capacity optimization model as the optimization objectives. Combined with the equipment capacity configuration constraints, it solves the capacity configuration scheme of each energy device. It addresses the shortcomings of existing capacity optimization that do not take into account economic efficiency, resource utilization efficiency, and power supply reliability, and the lack of constraints that leads to configuration schemes that are out of touch with reality. It achieves multi-objective collaborative optimization within the scope of compliance constraints, improves the wind and solar energy absorption rate, ensures the stability of the system's power supply, and controls investment costs.

[0016] S4. The lower-level scheduling optimization model takes minimizing the overall volatility of the electrolytic cell array as the optimization objective and sets system operation constraints to solve the optimal scheduling scheme under a given capacity configuration scheme; wherein, the overall volatility index is obtained by weighted calculation of multiple factors characterizing the start-up and shutdown characteristics and the stability of operating condition transitions of the electrolytic cell array.

[0017] Among them, the overall volatility of the electrolyzer array is a quantitative indicator characterizing the overall operational volatility of the electrolyzer array; start-up and shutdown characteristics are the operational features of the electrolyzer during cold start, hot start, shutdown, and standby; operating condition transition stability is the smoothness of the electrolyzer switching between shutdown, standby, fluctuating operation, and rated operation conditions; system operation constraints are the limiting conditions for equipment operating power, energy storage status, electrolyzer operating condition transitions, and start-up and shutdown actions; and the optimal scheduling scheme is the optimal real-time operation strategy of the system under a given capacity configuration. In this embodiment, S4 takes the minimum overall volatility of the electrolyzer array in the lower-level scheduling optimization model as the optimization objective, and solves the optimal scheduling scheme under a given capacity configuration by combining system operation constraints. The overall volatility of the electrolyzer array is obtained by weighted calculation of multiple factors characterizing start-up and shutdown characteristics and operating condition transition stability. This invention addresses the shortcomings of existing electrolyzer volatility quantification methods, such as single-dimensional characterization, failure to consider the coupling characteristics of multiple electrolyzers, and strong subjectivity of the penalty cost mechanism. It achieves accurate quantification of electrolyzer array volatility, reduces electrolyzer equipment losses, and improves the stability of hydrogen source supply.

[0018] S5. The upper-level capacity optimization model is solved using the first optimization algorithm to obtain a non-dominated solution set including multiple capacity configuration schemes. For each capacity configuration scheme in the non-dominated solution set, the lower-level scheduling optimization model is solved using the second optimization algorithm to obtain the corresponding preferred scheduling scheme and operating cost.

[0019] In this embodiment, the first optimization algorithm is an intelligent solution algorithm adapted to the upper-level multi-objective capacity optimization problem, and the second optimization algorithm is an intelligent solution algorithm adapted to the lower-level single-objective scheduling optimization problem. The non-dominated solution set is the Pareto optimal capacity configuration scheme set obtained from the upper-level multi-objective optimization, and the running cost is the real-time running cost generated by the system executing the scheduling scheme. In S5 of this embodiment, the first optimization algorithm is used to solve the upper-level capacity optimization model to obtain a non-dominated solution set containing multiple capacity configuration schemes. Then, for each capacity configuration scheme in the non-dominated solution set, the second optimization algorithm is used to solve the lower-level scheduling optimization model to obtain the optimal scheduling scheme and running cost of the corresponding capacity scheme. In view of the defects of high difficulty in solving multi-objective capacity optimization and low efficiency in solving the dual-layer coupled model, the efficient calculation of the dual-layer model is achieved through the layered collaborative solution of the two algorithms, completing the bidirectional feedback between the upper-level capacity configuration and the lower-level operation scheduling, and obtaining a comprehensive set of optimal solutions for system capacity configuration and scheduling.

[0020] S6. Based on all capacity configuration schemes and their corresponding preferred scheduling schemes, select the target capacity configuration scheme and its corresponding preferred scheduling scheme from the non-dominated solution set using a multi-attribute decision method.

[0021] Among them, the multi-attribute decision method is a decision method that selects the comprehensive optimal solution from the multi-objective optimal solution set. The target capacity configuration scheme is the finally selected system optimal equipment installed capacity scheme, and the target preferred scheduling scheme is the system optimal operation scheduling scheme that matches the target capacity configuration. In the S6 embodiment of the present invention, based on all capacity configuration schemes and corresponding preferred scheduling schemes, the multi-attribute decision method is used to select the target capacity configuration scheme and the corresponding preferred scheduling scheme from the non-dominated solution set. In view of the defect that the optimal scheme in the multi-objective optimization solution set is difficult to select intuitively, the comprehensive optimal system configuration and scheduling scheme is determined through scientific decision-making. In the end, the efficient consumption of wind and solar power, stable supply of hydrogen source and optimal hydrogen production cost of large-scale wind, solar and energy storage hydrogen production system are achieved, providing technical support for the large-scale application of green hydrogen.

[0022] Based on the above embodiments, as a preferred implementation, in step S1, the method for obtaining the wind and solar power output scenario and the hydrogen load scenario includes: Historical wind and solar resource data and historical hydrogen consumption load data are collected. The k-means clustering analysis algorithm is used to extract multiple cluster centers from the historical wind and solar resource data and the historical hydrogen consumption load data, respectively, as the wind and solar power output scenario and the hydrogen consumption load scenario.

[0023] Among them, historical wind and solar resource data are time-series data collected on-site at large-scale wind, solar, energy storage, and hydrogen production system application sites, reflecting the long-term output patterns of photovoltaic and wind power; historical hydrogen load data are energy consumption data collected over a long period at the same site, reflecting the time-series changes in hydrogen demand; the k-means clustering analysis algorithm is an unsupervised data processing method used to extract typical operational characteristics from massive time-series data; cluster centers are typical data points obtained through clustering calculations that can characterize the overall distribution characteristics of a class of data; wind and solar output scenarios are standardized scenarios that can reflect the typical power generation characteristics of regional wind and solar resources; and hydrogen load scenarios are standardized scenarios that can reflect the typical consumption patterns of regional hydrogen demand. This step... By collecting historical wind and solar resource data and historical hydrogen load data of the planning site, original data matrices are constructed and k-means clustering analysis algorithm is used to cluster the data, extracting the corresponding cluster centers. The cluster centers are used as wind and solar power output scenarios and hydrogen load scenarios, solving the problems of data redundancy, large computational load, and difficulty in reflecting typical system operating conditions caused by directly using full original time series data for optimization calculations in existing technologies. This achieves the typicalization and dimensionality reduction of wind and solar resource and hydrogen load data, providing stable, reliable, and representative input conditions for subsequent system model building and two-level optimization calculations, improving the computational efficiency of the optimization model and the degree to which the results fit the actual operating conditions.

[0024] Specifically, based on historical landscape data and historical hydrogen consumption data collected on-site, an original meteorological data matrix is ​​constructed. A 0= ( a ij ) n1×2 and raw energy consumption data matrix B 0 = ( b ij ) s×r1 ,in, a ij Representing the i The first form of energy j One meteorological parameter, b ij Representing the i The first form of energy j Energy consumption data, n 1 indicates the number of samples in the historical landscape data, that is, the length of the time series of landscape resource data collected. s This indicates the number of samples in the historical hydrogen load data, i.e., the length of the time series of hydrogen load data collected. r 1 represents the number of hydrogen usage scenarios, i.e., the number of different hydrogen users or hydrogen usage conditions served by the system; a typical meteorological data matrix is ​​obtained using the k-means clustering algorithm. A Compared with typical energy consumption matrix B .

[0025] Based on the above embodiments, as a preferred implementation, in step S1, the power subsystem includes a photovoltaic array, a wind power system, and a battery; the hydrogen subsystem includes a hydrogen storage tank of a first pressure level and a hydrogen storage tank of a second pressure level, wherein the first pressure level is higher than the second pressure level; the electrolyzer array includes multiple electrolyzers arranged in an array; and the large-scale wind-solar-storage-hydrogen production system model also includes a hydrogen metallurgy unit and a hydrogen refueling station.

[0026] Among them, the power subsystem is the core unit for realizing the production, storage and consumption of electricity in the large-scale wind, solar and energy storage hydrogen production system; the hydrogen subsystem is the key unit for realizing hydrogen storage and pressure matching; the electrolyzer array is the core device for realizing the efficient conversion of electrical energy into hydrogen energy; the hydrogen metallurgy unit is the hydrogen-using terminal unit for carrying out metallurgical production using green hydrogen; and the hydrogen refueling station is the hydrogen-using terminal unit for providing refueling services for hydrogen-powered vehicles.

[0027] This invention, based on pre-acquired wind and solar power output scenarios and hydrogen load scenarios, constructs a large-scale wind-solar-storage-hydrogen production system model. The power subsystem comprises a photovoltaic array, a wind power system, and batteries. The photovoltaic array and wind power system, as renewable energy generation units, can provide clean electricity to the system according to wind and solar resource conditions. The batteries, as energy storage units, can store electricity when wind and solar power output is abundant and release electricity when wind and solar power output is insufficient, achieving peak shaving and valley filling of electricity. The hydrogen subsystem includes hydrogen storage tanks of a first pressure level and a second pressure level, with the first pressure level being higher than the second. The low-pressure hydrogen storage tank can receive hydrogen produced by the electrolyzer array and perform preliminary storage. The high-pressure hydrogen storage tank can compress the low-pressure hydrogen to a high-pressure state through a pressurization device to meet the pressure requirements of different end-use hydrogen scenarios. The electrolyzer array consists of multiple units arranged in an array. The electrolyzer, powered by the power subsystem, decomposes water into hydrogen and oxygen, converting electrical energy into hydrogen energy. Simultaneously, the large-scale wind-solar-storage-hydrogen production system model includes two types of hydrogen-using terminals: a hydrogen metallurgy unit and a hydrogen refueling station. The hydrogen metallurgy unit utilizes the green hydrogen produced by the system for reduction smelting, while the hydrogen refueling station directly receives hydrogen from high-pressure hydrogen storage tanks to provide refueling services for vehicles. Addressing the shortcomings of existing wind-solar-storage-hydrogen production system models that do not adequately consider the coordinated matching of hydrogen storage units at different pressure levels and suffer from insufficient system adaptability due to the limited types of hydrogen-using terminals, this model constructs a complete system model encompassing multiple energy forms, multiple pressure levels of hydrogen storage, and multiple terminal hydrogen-using scenarios. This provides a comprehensive and practical system carrier for subsequent dual-layer capacity configuration optimization, ensuring that the optimization results can simultaneously adapt to the needs of different hydrogen-using terminals, thereby improving the overall practicality and application scenario adaptability of the system.

[0028] Based on the above embodiments, as a preferred implementation, in step S3, the annualized total cost includes system investment cost, system maintenance cost, and system operating cost.

[0029] The curtailment rate refers to the proportion of photovoltaic and wind power generation that is not effectively utilized due to grid transmission limitations, insufficient absorption, or lagging energy storage.

[0030] The energy supply reliability is the ratio of the sum of the number of hours that meet the electricity load and the number of hours that do not meet the hydrogen load to the sum of the total number of hours of electricity load and the total number of hours of hydrogen load.

[0031] Among them, annualized total cost is the core indicator for measuring the economic efficiency of a large-scale wind-solar-storage-hydrogen production system throughout its entire life cycle; curtailment rate is a key indicator for characterizing the utilization efficiency of wind and solar energy; and power supply reliability is an important indicator for reflecting the stability of the system's power and hydrogen supply. The upper-level capacity optimization model takes minimizing annualized total cost, minimizing curtailment rate, and maximizing power supply reliability as its optimization objectives. Specifically, the minimized annualized total cost is achieved using the following formula: Min TC = Cinv + C ma + C op in, TC This represents the annualized total cost. C inv For system investment costs, C ma for System maintenance costs C op The system operating cost is calculated using the following formula:

[0032]

[0033] in, I It is a collection of hydrogen production and storage equipment, including photovoltaic arrays, wind turbines, electrolyzer arrays, fuel cells, batteries, hydrogen storage tanks, heat exchangers, and heat accumulators; i I For the first in the device set i This type of equipment; r The loss rate is 8% in this embodiment of the invention; x i For the first i The service life of this type of equipment; c i,inv The unit price of each piece of equipment in the equipment set; P i,in The design capacity of each device; k i This refers to the maintenance cost coefficient for each piece of equipment in the equipment set. C op It is calculated by the lower-level scheduling plan model and fed back to the upper-level optimization model.

[0034] The curtailment rate is calculated using the following formula:

[0035] in, ECR For the curtailment rate, P loss,t for t Unused electricity at any given time, kWh; P pv,t for t Photovoltaic power generation at any given time, in kWh; P wt,t for t Wind turbine power generation at any given time, in kWh; curtailment rate is the proportion of photovoltaic and wind power generation that is not effectively utilized due to grid transmission limitations, insufficient absorption, or lagging energy storage.

[0036] Furthermore, the power supply reliability corresponding to the power supply unreliability rate is expressed by the formula: HLR = ( H Elock + H H2lock ) / ( H E + H H2 ) In the above formula, HLR The system does not meet the ratio of electricity to hydrogen load hours. H Elock To meet the required electricity load hours, H H2lock To meet the gas load hours, H E Total hours of electrical load H H2 This represents the total number of hours of hydrogen load.

[0037] By clearly defining and formulating the three core optimization objectives mentioned above, the problems of vague optimization objectives, unclear indicator definitions, and inability to quantify constraints in existing technologies, which lead to configuration schemes deviating from actual working conditions, are solved. This makes the upper-level capacity optimization objectives more accurate, quantifiable, and calculable, providing a stable and reliable objective guide and calculation basis for subsequent solutions to reasonable, economical, and efficient equipment capacity configuration schemes.

[0038] Based on the above embodiments, as a preferred implementation, in S3, the equipment capacity configuration constraints include: the curtailment rate not exceeding 8%, the energy supply reliability not less than 90%, the installed capacity of each energy device being between its preset lower and upper capacity limits, and the annualized total cost not exceeding the maximum allowable investment cost.

[0039] In this embodiment, the equipment capacity configuration constraint is a boundary condition set to ensure the safe, economical, and efficient operation of a large-scale wind, solar, energy storage, and hydrogen production system, used to limit the feasible solution space of the upper-level capacity optimization model. In the upper-level capacity optimization model, the equipment capacity configuration constraint includes curtailment rate constraints, energy supply reliability constraints, equipment installed capacity constraints, and annualized total cost constraints. The curtailment rate constraint is set to a curtailment rate not exceeding 8%. ECR <8%), this constraint limits the minimum level of wind and solar energy consumption, avoids resource waste caused by excessive power curtailment, and ensures the efficient use of renewable energy; the power supply reliability constraint is set at no less than 90% (i.e., HLR<10%), this constraint limits the minimum guarantee level of the system's power and hydrogen supply services, ensuring that the system can stably meet the vast majority of power and hydrogen load demands and reduce the risk of load shortfall; the installed capacity constraints for each energy device are that the installed capacity of photovoltaic arrays, wind power systems, batteries, hydrogen storage tanks, electrolyzer arrays, etc., are respectively between their preset lower and upper capacity limits (i.e., ,in, P in,i For the first i Installed capacity of this type of equipment P min in,i For the first i Lower limit of installed capacity for this type of equipment Pmax in,t For the first i The upper limit of installed capacity for each type of equipment is defined by the equipment's technical characteristics, site resources, and industry-standard configuration range. This limit restricts the reasonable configuration range for each piece of equipment, preventing situations where insufficient capacity fails to meet load demand or excessive redundancy leads to wasted investment. The annualized total cost constraint ensures that the system's annualized total cost does not exceed the preset maximum allowable investment cost. TC < TC max This constraint sets a clear upper limit for the system configuration scheme from the perspective of project economics, ensuring that the scheme is within the investment and operating cost range that the project entity can bear. By setting the above-mentioned multi-dimensional and multi-level equipment capacity configuration constraints, the problems of existing technologies such as the lack of clear boundary conditions in optimization models, the easy occurrence of resource waste, insufficient reliability, or exceeding the budget in configuration schemes are solved. This makes the solution process of the upper-level capacity optimization model more in line with the technical, economic, and resource constraints of actual projects, effectively narrowing the range of non-dominated solution sets and improving the feasibility, practicality, and implementability of capacity configuration schemes.

[0040] Based on the above embodiments, as a preferred implementation, in step S4, as follows: Figure 2 As shown, the multiple factors include cold start interference factor, hot start interference factor, start stability factor, shutdown-standby interference factor, fluctuation stability factor, and fluctuation continuity factor.

[0041] The weighting coefficients are calculated using the CRITIC method, an improved indicator weighting method based on indicator correlation and improved upon the maximum information coefficient, including: The maximum information coefficient is used to measure the nonlinear correlation between factors to determine the conflict, and the standard deviation of each factor is combined to determine the information carrying capacity of each factor, and then the weight of each factor is obtained by normalization.

[0042] Among them, the overall volatility of the electrolytic cell array is the core indicator used to quantify the overall volatility of multiple electrolytic cells operating in a coupled manner. Cold start interference factor, hot start interference factor, start-up stability factor, shutdown-standby interference factor, volatility stability factor, and volatility continuity factor are sub-evaluation indicators characterizing the start-up and shutdown characteristics and operational condition transition stability of the electrolytic cells, respectively. Weighting coefficients are quantitative coefficients reflecting the contribution of each factor to the overall volatility. The Maximum Information Coefficient (MIC) is a correlation metric that can capture linear and nonlinear relationships between variables. CRITIC (Criteria Importance Through Intercriteria) is also relevant. The Correlation (indicator weight determination based on indicator correlation) method is an objective weight assignment method based on indicator correlation and dispersion. The system operation constraints are power, energy storage, and operating condition switching boundary conditions set to ensure the safe and stable operation of the equipment. The lower-level scheduling optimization model takes minimizing the comprehensive volatility of the electrolytic cell array as the optimization objective and sets system operation constraints to solve the optimal scheduling scheme under a given capacity configuration. The comprehensive volatility index is calculated by weighting cold start interference factor, hot start interference factor, start stability factor, shutdown-standby interference factor, volatility stability factor, and volatility continuity factor, which characterize the start-up and shutdown characteristics and operating condition switching stability of the electrolytic cell array. The weighting coefficients are calculated using the CRITIC method, which is an improvement on the maximum information coefficient. The maximum information coefficient measures the nonlinear correlation between factors to determine the index conflict, and the standard deviation of each factor is combined to determine the index dispersion. Then, the information carrying capacity of each factor is calculated based on the conflict and dispersion. Finally, the information carrying capacity is normalized to obtain the objective weight of each factor. This weight is objectively calculated from the data, which can avoid the subjective bias caused by manual setting. At the same time, the lower-level scheduling model sets system operation constraints such as component operating power constraints, electrolyzer operating power constraints, upper and lower bound constraints of hydrogen storage tank operation, electrolyzer operating status and transformation logic constraints, and start-up and shutdown action constraints.

[0043] By constructing a multi-dimensional volatility factor and employing the MIC-modified CRITIC method for objective weighting, combined with comprehensive system operation constraints, this approach solves the problems of existing technologies that only use a single-dimensional index to characterize volatility, cannot reflect the coupled operation characteristics of multiple electrolytic cells, have strong subjective penalty costs, and whose incomplete constraints lead to discrepancies between scheduling results and actual operating conditions. This achieves accurate, comprehensive, and objective quantification of the volatility of electrolytic cell array operations, providing a scientifically reliable objective function and feasible solution space for lower-level scheduling optimization. This improves the stability, rationality, and equipment friendliness of scheduling schemes, effectively reducing electrolytic cell losses and extending equipment lifespan. Based on the above embodiments, as a preferred implementation, the cold start interference factor is the product of the normalized value of the average cold start intensity of a single electrolytic cell and the cold start frequency.

[0044] Specifically, in this embodiment, considering the start-up and shutdown characteristics of the electrolytic cell, the following measures are taken: n In an array of electrolytic cells, each electrolytic cell is classified according to its time-series operating condition characteristics. Each electrolytic cell has four operating conditions, including shutdown conditions. D Standby mode S Fluctuating operating conditions F Rated operating conditions R The electrolytic cell's air-condition switching action is divided into two categories: starting the motor as... Y , as a motor vehicle Z The power-on process is divided into cold start and warm start. Switching from shutdown to standby mode requires a warm start, which takes approximately [time missing]. T HS Switching from standby mode to fluctuating operation mode requires a cold start, which takes approximately [time missing]. T CS .

[0045] right n The statistical parameters of the macroscopic operating characteristics of the Taiwan electrolytic cell were standardized and normalized.

[0046] For the i Calculate the cold start intensity of a single electrolytic cell and compile statistics. T Number of cold starts within a time period Ncs (i) is defined as the electrolytic cell being in shutdown condition. D Switch to standby mode S Fluctuating operating conditions F Total number of switching times under rated operating condition R; statistics T Cold start intensity within a given time period, and establish a cold start intensity matrix. P CS_1 , P CS_2 , ..., P CS_Ncs ] 1×Ncs ( k =1... Ncs ( i )),in P CS_k ( i ) is defined as the first k Starting power during cold start (electrolytic cell starting from 0 to power of) P CS_k ( i ), k =1... Ncs ( i )).

[0047] Calculate the average intensity matrix for cold start of the electrolytic cell array: P CS_mean =[ P CS_mean (1), P CS_mean (2),..., P CS_mean (n)] Among them, the i Average cold start intensity of TECHNOLOGY cell P CS_mean ( i )for: P CS_mean ( i )= ( Ncs ( i )=0, P CS_mean ( i )=0) The average intensity matrix of the electrolytic cell array during cold start is standardized. i The normalized average cold start intensity value of the TSMC electrolytic cell is:

[0048] in, This represents the minimum average cold start intensity of all electrolytic cells in the array. This represents the maximum value of the average cold start intensity of all electrolytic cells in the array.

[0049] Calculate the cold start frequency of a single unit. f CS Defined as the ratio of cold start time to the total operating time of the electrolyzer: (0 ≤ f CS ( i ) ≤1) in T total ( i )for T Total operating time of the electrolytic cell during the period For the first i The cold start frequency of the TECH solvent cell For the first i The total cold start time of the TECH solvent cell during period T.

[0050] Establish a single-unit cold start interference factor, and measure it using the average cold start intensity and cold start frequency: S 1( i )= P CS_mean norm ( i )× f CS ( i ) in, S 1( i ) is the first i Cold start interference factor of the TSMC electrolytic cell.

[0051] The hot start interference factor is the product of the normalized value of the average hot start intensity of a single electrolytic cell and the hot start frequency.

[0052] Calculate the thermal start-up intensity of a single electrolytic cell and statistically analyze it. T Number of hot starts within a time period N HS ( i ), defined as the electrolytic cell in standby mode. S Switch to fluctuating operating conditions F Rated operating conditions R Total number of switches; statistics T The intensity of hot start within a given time period is determined, and a hot start intensity matrix is ​​established. [ P HS_1 , P HS_2 ,..., P HS_NHS ] 1×NHS ( j =1... N HS ( i )) in, P HS_j Defined as the first j Starting power during secondary hot start (electrolytic cell from standby power) P 0 Startup to P HS_j ( i ), j =1... N HS ( i )); Calculate the first i Taiwan Electrolytic Cell Array Hot Start-up Average Intensity Matrix P HS_mean ( i ): P HS_mean =[ P HS_mean (1), P HS_mean (2),..., P HS_mean (n)] in, P HS_mean ( i )for: P HS_mean ( i )= ( N HS ( i ) = 0, P HS_mean ( i ) = 0) in, P HS_mean Let be the average intensity matrix of the electrolytic cell array during hot start-up; after standardizing the average intensity matrix of the electrolytic cell array during hot start-up, we obtain:

[0053] No. i Hot start frequency of TECHNOLOGY Defined as the ratio of hot start-up time to the total operating time of the electrolyzer: (0 ≤ f HS ( i ) ≤1) in, For the first i Taiwan Electrolytic Cells T Total warm-up time within the time period For the first i Taiwan Electrolytic Cells T Total running time within the time period; establish the first i Single-unit hot-start interference factor of an electrolytic cell: S 2( i ) = P HS_mean_norm ( i )× f HS ( i ) The startup stability factor is the ratio of the total startup time to the total operating time of the electrolyzer, specifically: S 3( i ) = fe ( i ) in, S 3( i ) is the first i The start-up stability factor of the TECH solvent cell. f e ( i ) is the first i The total start-up frequency of an electrolytic cell is defined as the ratio of the total start-up time to the total operating time, and is numerically equal to the sum of the cold start frequency and the hot start frequency. Start-up frequency f e Defined as the ratio of the total start-up time to the total running time of the electrolyzer: f e ( i ) = f CS ( i ) + f HS ( i ) (0 ≤ f e ( i ) ≤1) The power-off / standby interference factor is calculated based on the sum of the number of power-offs and standbys, the average of the normalized power-off intensity value and the normalized standby intensity value, and the maximum allowable number of state transitions.

[0054] Specifically, calculate the shutdown intensity of a single electrolytic cell and statistically analyze... T Number of times the device was shut down during the period N D Defined as the electrolytic cell operating under standby conditions S、 Fluctuating operating conditions F、 Rated operating conditions R Switch to shutdown mode D's total Number of times; statistics T The power-off intensity during the time period is determined, and a power-off intensity matrix is ​​established. P D_1 , P D_2 ,..., P D_ND ] 1×ND ,in, P D_m Defined as the first m Power consumption during the first shutdown; Calculate the average intensity matrix of the electrolytic cell array when it is turned off. P D_mean =[ P D_mean (1), PD_mean (2),..., P D_mean (n)], where the nth i Average shutdown intensity of TSMC electrolytic cell P D_mean ( i )for: P D_mean ( i )= ( N D ( i ) = 0, P D_mean ( i ) = 0) in, N D ( i ) is the first i Taiwan Electrolytic Cells T Number of times the device was shut down within a given time period For the first i Taiwan Electrolytic Cell No. m The power during the second shutdown is used to standardize the average shutdown intensity matrix of the electrolytic cell array, as shown in the above embodiment.

[0055] T Number of standby times during the period N S Defined as the electrolytic cell operating under fluctuating conditions. F、 Switching from rated operating condition R to standby condition S total frequency; T Standby intensity matrix during the time period [ P S_1 , P S_2 ,..., P S_ND ] 1×ND ,in P S_i Defined as the first z The power during standby (i.e., from) P S_i Up to standby power P 0); Calculate the standby average intensity matrix of the electrolytic cell array P S_mean =[ P S_mean (1), P S_mean (2),..., P S_mean (n)], where the nth iAverage standby intensity of TECHNOLOGY cell P S_mean ( i )for: P S_mean ( i )= ( N S ( i ) = 0, P S_mean ( i ) = 0) in, For the first i Taiwan Electrolytic Cell No. z Power consumption during standby; N S ( i ) is the first i Taiwan Electrolytic Cells T The number of standby cycles within a time period; the average standby intensity matrix of the electrolytic cell array is standardized using the same method as above.

[0056] No. i Standby interference factor of a single electrolytic cell during shutdown for:

[0057] in, N max for T The maximum number of times the electrolytic cell status can be changed within a given time period; For the first i Standardized value of average shutdown strength of TSMC electrolytic cell; For the first i Standardized value of the standby average intensity of the Taiwan electrolytic cell.

[0058] The fluctuation stability factor is the sum of the products of the standard deviation of power fluctuation in each time period of the electrolyzer under continuous fluctuation conditions and the fluctuation frequency of the corresponding time period.

[0059] The fluctuation continuity factor is the ratio of the sum of fluctuation frequencies to the sum of standby frequency and startup frequency.

[0060] Specifically, standby frequency f S Defined as the ratio of the total standby time of the electrolyzer to the total operating time of the electrolyzer: (0 ≤ f S ( i ) ≤1) Number of time periods under continuous fluctuation conditionsk That is, k A period of continuous fluctuation, During the time period, the electrolytic cell was operating under continuous fluctuation conditions, and the calculations were performed for the first time. i Taiwan Electrolytic Cell No. l Standard deviation of power fluctuation under continuous fluctuation conditions σ i Power fluctuation standard deviation matrix σ= [ σ 1, σ 2 ,...,σ l ,...,σ k ]: σ =

[0061] in, Indicates the first l Duration of the fluctuating operating condition; For the first i Taiwan electrolytic cell in the first l Within the fluctuating operating conditions t Operating power at any given time; For the first i Taiwan electrolytic cell in the first l Average operating power within the fluctuating operating conditions; for the power fluctuation standard deviation matrix [ σ 1, σ 2 ,...,σ l ,...,σ k Standardize:

[0062] in, For the first i Taiwan electrolytic cell in the first l Standardized value of power fluctuation standard deviation under segment fluctuation conditions; Calculation of fluctuation frequency of each segment of the electrolytic cell under continuous fluctuation conditions within time period T. f F,l， Fluctuation frequency matrix [ f F , 1 f F , 2 ,...,f F,l ,...,f F,k ] Defined as the first l The ratio of the fluctuating operating time of the electrolytic cell to the total operating time of the electrolytic cell: (0 ≤ fF ( i ) ≤1) in, For the first i Taiwan electrolytic cell in the first l Running time under fluctuating operating conditions; For the first i Taiwan electrolytic cell in the first l The frequency of the fluctuating operating condition; the first i Taiwan electrolytic cell in the first l Segment volatility stability factor S 5( i )for: S 5( i )=

[0063] No. i Taiwan electrolytic cell in the first l Segment fluctuation continuity factor for:

[0064] Furthermore, such as Figure 3 As shown, this embodiment of the invention utilizes the MIC-modified CRITIC method for weight assignment, for cold start interference factor, hot start interference factor, start stability factor, shutdown-standby interference factor, fluctuation stability factor, and fluctuation continuity factor (…). S 1. S 2、....、 S 6) Six factor indicators, using standard deviation to represent the variation and fluctuation of the values ​​of each indicator. The larger the standard deviation, the greater the numerical variation of the indicator, the more information it can reflect, and the stronger the evaluation strength of the indicator itself. More weight should be assigned to the indicator.

[0065] That is, for those by n An array of electrolytic cells consisting of multiple electrolytic cells:

[0066]

[0067] in, As a factor j standard deviation For the first i Taiwan Electrolytic Cell's first j The index value of each volatility factor, j For the first j The number of each volatility factor indicator ( j =1,2,3,4,5,6). For all electrolytic cellsj The average value of each volatility factor indicator.

[0068] In the CRITIC method, conflict reflects the degree of correlation between different indicators. A smaller conflict value indicates a significant positive correlation. Traditional CRITIC methods typically use the Pearson correlation coefficient to calculate the correlation between different indicators. However, the Pearson correlation coefficient is generally used to capture linear associations and is almost incapable of identifying nonlinear relationships common in machine fluctuation scenarios, leading to distorted conflict calculations. The Maximum Information Coefficient (MIC) is an indicator that measures the strength of any association between two variables, regardless of whether the relationship is linear, nonlinear, or non-monotonic. As long as there is a data-supported association between the variables, the MIC can provide a result matching the strength. Its value ranges from [0,1], with a value closer to 1 indicating a stronger association. Using the MIC instead of the Pearson correlation coefficient to calculate conflict allows conflict to more accurately reflect the true nonlinear association between indicators, ultimately making the weight allocation more consistent with data patterns. The definition of MIC is as follows: if there is a relationship between two variables, the two variables can be grouped into a finite set D. A grid is drawn on the scatter plot of set D. These grids can divide the data in the scatter plot. Based on the distribution of points in the grid, the probability distribution under this division method can be obtained. Entropy and mutual information are calculated through the probability distribution. The resolution of the grid is gradually increased. The maximum mutual information value under all resolutions is obtained by traversing and normalizing the calculation.

[0069] The formula for calculating MIC is: (m×n <B) in, S a , S b Here are two volatility factor indices for which the correlation strength is to be calculated. m and n These represent the number of rows and columns in the grid, respectively, and B is the maximum number of grid cells. I ( S a ; S b ) is a variable S a With variables S b The mutual information is calculated using the following formula:

[0070] in, p ( s i ; s j )for S i andS j Joint probability distribution p ( s i )for S a Pick s i Marginal probability at time p ( s j )for S b Pick s j The marginal probability at that time.

[0071] The MIC correlation matrix is ​​established based on the MIC calculation results to replace the Pearson correlation coefficient matrix in the original method, that is:

[0072] Among them, matrix elements For the first j The first indicator and the first k The maximum information coefficient of each indicator is used to define the conflict by replacing the Pearson correlation coefficient with the MIC value:

[0073] in, m This represents the total number of volatility factor indicators. As an indicator j The sum of the MIC values ​​of all other factors reflects the index. j Overall correlation with other factors.

[0074] No. j Information carrying capacity of each volatility factor indicator Calculated by the product of dispersion and conflict:

[0075] Normalization calculation j Weights of each factor : (j = 1,2,3...) Calculate the overall volatility index of the electrolytic cell array As the lower-level objective function:

[0076] in, For the first i Taiwan Electrolytic Cell's first j The index value of each volatility factor.

[0077] At the same time, the daily operating cost of the system is calculated based on the lower-level scheduling plan. C op The formula for feeding back to the upper-level optimization model is as follows: C op = C sell_H +C sell_E in, C sell_H For the revenue generated from the sale of hydrogen by the system, C sell_E The system generates revenue from selling electricity and hydrogen. C sell_H Calculate using the following formula: C sell_H = V sell_H (t) Price H_S (t) in, V sell_H ( t )for t The amount of hydrogen available for sale by the time-tracking system. Price H_S (t) is t The selling price of hydrogen at any given time.

[0078] Revenue from the sale of electricity C sell_E Calculate using the following formula: C sell_E = P_sell (t) Price E_S (t) in, P_sell ( t )for t The amount of electricity that the timekeeping system can sell. Price E_S (t) is t The selling price per unit of electricity at any given time.

[0079] Based on the above embodiments, this embodiment of the invention also establishes an operational constraint model for the wind-solar-storage coupled hydrogen production system in the lower-level scheduling model. Specifically, the power output of each output component in the system will not exceed its installed capacity during operation, that is:

[0080] in, P i (t ) is the first i Various types of equipment (photovoltaics, wind turbines, electrolyzers, batteries, hydrogen storage tanks, etc.) in t Always put in the effort; P i,in For the first i The design capacity of this type of equipment.

[0081] The operating power constraint of the electrolytic cell is:

[0082] in, P AE (t) represents the operating power of the electrolytic cell at time t, in kW; P AE,in The design capacity of the electrolytic cell is expressed in kW. w AE,min , w AE,max It is divided into the minimum operating power ratio and the maximum operating power ratio of the electrolytic cell.

[0083] The operating conditions and operating condition transformation logic constraints of the electrolytic cell are as follows: After the electrolytic cell is shut down, 1 h Restart time:

[0084] in, For the first i Taiwan Electrolytic Cells t The shutdown status variable at any given time (1 for shutdown, 0 for running / standby).

[0085] The constraint relationship between the electrolytic cell operating condition state variables and the start-up and shutdown action variables is as follows:

[0086]

[0087] in, For the first i Taiwan Electrolytic Cells t The rated operating condition state variable at any given time (1 indicates rated operating condition, 0 indicates non-rated operating condition). For the first i Taiwan Electrolytic Cells t The fluctuation of the operating condition state variable at any time (1 indicates that it is in the rated operating condition, 0 indicates otherwise); For the first i Taiwan Electrolytic Cells t The standby operating condition state variable at any given time (1 indicates that it is in the rated operating condition, and 0 indicates otherwise); For the first i Taiwan Electrolytic Cellst The shutdown status variable at any given time (1 indicates that the machine is in the rated operating condition, and 0 indicates otherwise). For the first i Taiwan Electrolytic Cells t The power-on action variable at time (1 for rated operation, 0 for otherwise); Zt(i): the value of the first... i Taiwan Electrolytic Cells t The shutdown action variable at any given time (1 indicates that the system is in rated operating condition, and 0 indicates otherwise).

[0088] No. i At any time, the TECH solvent cell t It can only be in one operating condition:

[0089] The operating constraints of the storage battery are:

[0090] SOC ES,min <SOC ES (t) <SOC ES,max 0≤ P ES,chr ( t )≤ P ES,chr,max 0≤ P ES,dis ( t )≤ P ES,dis,max μ chr ( t )+ μ dis ( t )≤1 in, ES ( t )yes t The battery stores energy at all times; SOC ES (t) is t The state of charge of the battery at all times; η loss_ES , η chr_ES , η dis_ES These are the battery self-discharge coefficient, charging efficiency, and discharging efficiency; P ES,chr ( t ), P ES,dis (t ) are respectively the batteries in t Constant charging and discharging power; SOC ES,max , SOC ES,min These are the upper and lower limits of the battery load state, respectively. P ES,chr,max , P ES,dis,max These are the maximum charging and discharging power of the battery, respectively. μ chr_ES ( t ), μ dis_ES ( t ) is a storage battery t Charge and discharge state variables at time 0-1.

[0091] The operating constraints of the hydrogen storage tank are:

[0092] SOC TK,min <SOC TK ( t ) <SOC TK,max P TK,min ≤P TK ( t ) ≤P TK,max in, m ( t )yes t The mass of gas stored in the hydrogen storage tank at all times; η loss_TK , η chr_TK , η dis_TK These are the hydrogen storage tank self-release coefficient, filling efficiency, and filling efficiency, respectively. P TK,chr ( t ), P TK,dis ( t ) are hydrogen storage tanks in t Constant inflation / deflation power; SOC TK,max , SOC TK,min These are the upper and lower limits of the hydrogen storage tank load status, respectively. SOC TK ( t )for t The hydrogen charge status of the hydrogen storage tank at all times; PTK,max , P TK,max These are the maximum and minimum permissible pressures of the hydrogen storage tank, respectively. P TK ( t For hydrogen storage tanks t Constant operational pressure; μ chr_TK (t), μ dis_TK (t) represents the hydrogen storage tank. t The state variables for inflation and deflation at time 0-1.

[0093] Based on the above embodiments, as a preferred implementation method, such as Figure 4 As shown, the first optimization algorithm is the Non-dominated Sorting Genetic Algorithm II (NSGA-II), and the second optimization algorithm is the Particle Swarm Optimization (PSO) algorithm.

[0094] S5 specifically includes: The upper-level capacity optimization model is solved using a non-dominated sorting genetic algorithm to obtain a non-dominated solution set that includes multiple capacity configuration schemes.

[0095] For each capacity configuration scheme in the non-dominated solution set, the particle swarm optimization algorithm is used to solve the lower-level scheduling optimization model to obtain the optimal scheduling scheme and operating cost corresponding to the capacity configuration scheme.

[0096] The operating cost is fed back to the upper-level capacity optimization model to determine the system operating cost in the annualized total cost.

[0097] In this embodiment, the specific operation steps of the NSGA-II algorithm are as follows: (1) Initialize the population by randomly generating a population of size . N， Initial population P 0, each individual in the population corresponds to a feasible solution. x Calculate each individual x Corresponding objective function f(x) And the fitness of individual populations.

[0098] (2) Perform non-dominated sorting for two individuals in the population. x i and x j If the following conditions are met: i ∈1,2,...,n All have f ( x i ) ≤ f ( x j ) i ∈1,2,..., n , making f (x i )< f ( x j ) Then it is called an individual x i Dominant Individual x j For a population size of N population P Each individual x i Calculate the number of its dominant individuals (how many individuals dominate it). n i and the set of individuals it governs S i If an individual x i Dominant Individual x j Then the individual x i Add to collection S i In the middle; if an individual x j Dominant Individual x i ,but n i = n i +1. All n i Individuals with a value of 0 are placed in the first non-dominated layer. F 1. Then, eliminate the individual's dominance over the remaining individuals and recalculate the set. S i Each individual in n j ,like n j If the value is 0, then add it to the next non-dominated layer. F 2. Repeat the above steps until all individual levels have been assigned.

[0099] (3) Crowding distance calculation: In order to perform selection and ranking among individuals with the same level, the NSGA-II algorithm introduces a crowding operator to measure the density of each individual in its non-dominated layer. For the non-dominated layer... F i Initialize the crowding distance for each individual. d i It is 0. For the first m There are several objective functions, which are then sorted according to their objective function values. After sorting, the crowding distance for boundary individuals is set to infinity, and the crowding distance for non-boundary individuals is calculated based on their adjacent objective function values.

[0100] in, For the first m In the n objective functions, after sorting them in ascending order of function values, the nth i +1 individual objective function values, f max m and f min m They are the first m The objective function is defined as the maximum and minimum values ​​at this layer. By selecting individuals with larger crowding distances, the calculation results can be distributed more evenly in space, thus better maintaining population diversity.

[0101] (4) Selection, crossover, and mutation: The selection, crossover, and mutation strategies are largely similar to those of ordinary genetic algorithms. Based on non-dominated sorting and crowding, a tournament selection strategy is used to select the population. After selection, new offspring are generated through crossover and mutation operations. The crossover operation uses simulated binary crossover, and the mutation operation uses polynomial mutation.

[0102] (5) Elite Preservation Strategy: This strategy ensures that superior individuals from the parent generation are directly passed on to the offspring, preventing the loss of some excellent solutions. Specifically, the parent and offspring populations are first combined into a single population. R i First, based on the Pareto gradation, the entire layer is placed into the new parent population from low to high. C i+1 Until a certain layer appears C i+1 Exceeding population size N Then, based on the crowding distance of individuals in that layer, the population is filled from largest to smallest. C i+1 Until the population size is reached N .

[0103] Particle swarm optimization (PSO) abstracts each potential solution to an optimization problem as a "particle," and all particles form a "swarm." Each particle moves in the solution space, dynamically adjusting its position and velocity through individual experience (the optimal solution found in the particle's own history) and swarm experience (the optimal solution found in the history of the entire swarm). Through continuous iteration, it gradually converges towards the global optimum. The state of each particle is described by its position and velocity. The specific algorithm flow is as follows: First, randomly generate... N The initial position and initial velocity of each particle are determined; secondly, based on the optimization objective function, the fitness of each particle is calculated, and if the current particle's fitness is better than its historical best, then the particle is considered fit. pbest Then update the current position to pbest Again, in all particles pbest In the process, the position with the best fitness is selected as the global optimum. gbest The velocity and position of each particle are updated according to the above formula; finally, if the termination condition is met (such as reaching the maximum number of iterations), the process continues. gbest If the fitness is less than the threshold, then output gbest Otherwise, repeat the iteration. The specific steps are as follows: Suppose the solution space of the optimization problem is D Wei, then the first i The positions of the particles are: x i = ( x i1 ,x i2 ,...,x iD ) The speed is: v i = ( v i1 ,v i2 ,...,v iD ) The particle's velocity is updated using the following formula: v id ( t+1 ) =wv id ( t ) +c 1 r 1( pbest id x id ( t )) +c 2r 2( gbest d x id ( t )) In the formula, t This represents the current iteration number; w Inertial weights; c 1 ,c 2 represents the learning factor, which controls the influence of individual experience and group experience respectively; r 1 ,r 2 is a random number in the interval [0,1]. pbest id For the first i The particle in the first d The optimal position of an individual in dimension; gbest d为 The group in the d The globally optimal position of the dimension.

[0104] The particle's position is adjusted based on the updated velocity: x id ( t+1 ) =x id ( t ) +v id ( t+1 ) In the formula, x id ( t )for t Moment Particle i In the d Dimensional position; x id ( t+1 )for t+1 Moment Particle i In the d Dimensional position.

[0105] Based on the above embodiments, as a preferred implementation, in step S6, the multi-attribute decision-making method is an ideal solution method. The multi-attribute decision-making method is used to select a target capacity configuration scheme and a corresponding preferred scheduling scheme from the non-dominated solution set, including: Calculate the Euclidean distance from each capacity configuration scheme to the positive ideal solution and the negative ideal solution. Calculate the relative proximity of each capacity configuration scheme based on the Euclidean distance. Determine the capacity configuration scheme with the highest relative proximity and its corresponding preferred scheduling scheme as the target capacity configuration scheme and its corresponding preferred scheduling scheme.

[0106] Specifically, in this embodiment of the invention, based on the optimal solution set for system capacity configuration and scheduling, the evaluation criterion weight matrix is ​​calculated using TOPSIS (Technique for Order Preference by Similarity to an Ideal Solution), and the positive and negative ideal solutions for the wind-solar-storage coupled hydrogen production system are calculated. Based on the positive and negative ideal solutions, the target wind-solar-storage hydrogen production system is determined. Specifically, the TOPSIS method first needs to construct an evaluation system composed of multiple evaluation indicators according to the actual situation of the decision problem. Then, for each scheme, its value on each evaluation indicator needs to be calculated to form an evaluation matrix. Next, the evaluation matrix is ​​normalized to eliminate the influence of different indicator dimensions on the evaluation results. On this basis, the ideal solution and the negative ideal solution are determined, and then the relative distance (i.e., Euclidean distance) of each scheme to the ideal solution and the negative ideal solution is calculated. Finally, the schemes are ranked according to the relative distance of each scheme to the ideal solution; the scheme with the greater relative proximity has a better overall evaluation result. The formula for selecting the optimal solution using TOPSIS is as follows:

[0107]

[0108] (0) C i 1) In the formula, v+ j For the first j The positive ideal solution for each objective; v-j For the first j The negative ideal solution for each objective; S+ i The distance from the goal to the ideal solution; S- i The distance from the target to the negative ideal solution; C i For the first i The degree of closeness between the solutions is considered; the higher the degree of closeness, the better the target value.

[0109] In this embodiment of the invention, the relative proximity is expressed by a ratio to indicate the merits of the scheme. The larger the relative proximity, the closer the scheme is to the ideal distance and the farther it is from the negative ideal distance, and the better its overall benefits. Therefore, the relative proximity of the wind, solar, energy storage and hydrogen production systems can be sorted in descending order, and the wind, solar, energy storage and hydrogen production system with the largest relative proximity can be determined as the target wind, solar, energy storage and hydrogen production system.

[0110] Secondly, embodiments of the present invention provide a large-scale wind-solar-storage-hydrogen production dual-layer capacity configuration optimization device, based on the methods in the above embodiments, such as Figure 5 As shown, the device 500 includes: The scenario construction module 510 is used to construct a large-scale wind-solar-storage-hydrogen production system model based on the pre-acquired wind and solar power output scenario and hydrogen load scenario. The large-scale wind-solar-storage-hydrogen production system model includes a power subsystem, a hydrogen subsystem, and an electrolyzer array. The two-layer model construction module 520 is used to construct a two-layer capacity configuration optimization model, which includes an upper-layer capacity optimization model and a lower-layer scheduling optimization model. The upper-level optimization module 530 is used to optimize the power subsystem, hydrogen subsystem, and electrolyzer array with the minimum annualized total cost, minimum power curtailment rate, and maximum power supply reliability as the optimization objectives, and sets equipment capacity configuration constraints. It is used to solve the capacity configuration scheme of each energy device in the power subsystem, hydrogen subsystem, and electrolyzer array. The lower-level optimization module 540 is used to minimize the overall volatility of the electrolytic cell array as the optimization objective and to set system operation constraints to solve the optimal scheduling scheme under a given capacity configuration scheme; wherein, the overall volatility index is obtained by weighted calculation of multiple factors characterizing the start-up and shutdown characteristics and the stability of the operating condition transition of the electrolytic cell array; The two-layer solution module 550 is used to solve the upper-layer capacity optimization model using a first optimization algorithm to obtain a non-dominated solution set including multiple capacity configuration schemes. For each capacity configuration scheme in the non-dominated solution set, the lower-layer scheduling optimization model is solved using a second optimization algorithm to obtain the corresponding preferred scheduling scheme and operating cost. The decision output module 560 is used to select the target capacity configuration scheme and the corresponding preferred scheduling scheme from the non-dominated solution set using a multi-attribute decision method based on all capacity configuration schemes and the corresponding preferred scheduling schemes.

[0111] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the appended claims.

Claims

1. A method for optimizing the capacity configuration of a large-scale wind-solar-storage-hydrogen production dual-layer system, characterized in that, include: S1. Based on the pre-acquired wind and solar power output scenarios and hydrogen load scenarios, construct a large-scale wind, solar, storage and hydrogen production system model. The large-scale wind, solar, storage and hydrogen production system model includes a power subsystem, a hydrogen subsystem and an electrolyzer array. S2. Construct a two-layer capacity configuration optimization model, which includes an upper-layer capacity optimization model and a lower-layer scheduling optimization model. S3. The upper-level capacity optimization model takes the minimum annualized total cost, the minimum curtailment rate, and the maximum energy supply reliability as optimization objectives, and sets equipment capacity configuration constraints to solve the capacity configuration scheme of each energy device in the power subsystem, hydrogen subsystem, and electrolyzer array. S4. The lower-level scheduling optimization model takes minimizing the overall volatility of the electrolytic cell array as the optimization objective and sets system operation constraints to solve the optimal scheduling scheme under a given capacity configuration scheme; wherein, the overall volatility index is obtained by weighted calculation of multiple factors characterizing the start-up and shutdown characteristics and operating condition transition stability of the electrolytic cell array; S5. The upper-level capacity optimization model is solved using the first optimization algorithm to obtain a non-dominated solution set including multiple capacity configuration schemes. For each capacity configuration scheme in the non-dominated solution set, the lower-level scheduling optimization model is solved using the second optimization algorithm to obtain the corresponding preferred scheduling scheme and operating cost. S6. Based on all capacity configuration schemes and their corresponding preferred scheduling schemes, select the target capacity configuration scheme and its corresponding preferred scheduling scheme from the non-dominated solution set using a multi-attribute decision method.

2. The method for optimizing the capacity configuration of a large-scale wind-solar-storage-hydrogen production dual-layer system according to claim 1, characterized in that, In step S1, the methods for obtaining the wind and solar power output scenario and the hydrogen load scenario include: Historical wind and solar resource data and historical hydrogen consumption load data are collected. The k-means clustering analysis algorithm is used to extract multiple cluster centers from the historical wind and solar resource data and the historical hydrogen consumption load data, respectively, as the wind and solar power output scenario and the hydrogen consumption load scenario.

3. The method for optimizing the capacity configuration of a large-scale wind-solar-storage-hydrogen production dual-layer system according to claim 1, characterized in that, In S1, the power subsystem includes a photovoltaic array, a wind power system, and a battery; the hydrogen subsystem includes a hydrogen storage tank at a first pressure level and a hydrogen storage tank at a second pressure level, wherein the first pressure level is higher than the second pressure level; the electrolyzer array includes multiple electrolyzers arranged in an array; the large-scale wind-solar-storage-hydrogen production system model also includes a hydrogen metallurgy unit and a hydrogen refueling station.

4. The method for optimizing the capacity configuration of a large-scale wind-solar-storage-hydrogen production dual-layer system according to claim 1, characterized in that, In S3, the annualized total cost includes system investment cost, system maintenance cost, and system operating cost; The curtailment rate refers to the proportion of photovoltaic and wind power generation that is not effectively utilized due to grid transmission limitations, insufficient absorption, or lagging energy storage. The energy supply reliability is the ratio of the sum of the number of hours that meet the electricity load and the number of hours that do not meet the hydrogen load to the sum of the total number of hours of electricity load and the total number of hours of hydrogen load.

5. The method for optimizing the capacity configuration of a large-scale wind-solar-storage-hydrogen production dual-layer system according to claim 4, characterized in that, In S3, the equipment capacity configuration constraints include: the curtailment rate not exceeding 8%, the energy supply reliability not less than 90%, the installed capacity of each energy device being between its preset lower and upper capacity limits, and the annualized total cost not exceeding the maximum allowable investment cost.

6. The method for optimizing the capacity configuration of a large-scale wind-solar-storage-hydrogen production dual-layer system according to claim 1, characterized in that, In S4, the plurality of factors include a cold start interference factor, a hot start interference factor, a start stability factor, a shutdown-standby interference factor, a fluctuation stability factor, and a fluctuation continuity factor. The weighting coefficients are calculated using the CRITIC method, an improved indicator weighting method based on indicator correlation and improved upon the maximum information coefficient, including: The maximum information coefficient is used to measure the nonlinear correlation between factors to determine the conflict, and the standard deviation of each factor is combined to determine the information carrying capacity of each factor, and then the weight of each factor is obtained by normalization.

7. The method for optimizing the capacity configuration of a large-scale wind-solar-storage-hydrogen production dual-layer system according to claim 6, characterized in that, The cold start interference factor is the product of the normalized value of the average cold start intensity of a single electrolytic cell and the cold start frequency. The hot start interference factor is the product of the normalized value of the average hot start intensity of a single electrolytic cell and the hot start frequency. The startup stability factor is the ratio of the total startup time to the total running time of the electrolyzer. The power-off / standby interference factor is calculated based on the sum of the number of power-offs and standbys, the average of the normalized power-off intensity value and the normalized standby intensity value, and the maximum allowable number of state transitions. The fluctuation stability factor is the sum of the products of the standard deviation of the power fluctuation of the electrolyzer in each time period under continuous fluctuation conditions and the fluctuation frequency of the corresponding time period. The fluctuation continuity factor is the ratio of the sum of fluctuation frequencies to the sum of standby frequency and startup frequency.

8. The method for optimizing the capacity configuration of a large-scale wind-solar-storage-hydrogen production dual-layer system according to claim 1, characterized in that, The first optimization algorithm is a non-dominated sorting genetic algorithm, and the second optimization algorithm is a particle swarm optimization algorithm. S5 specifically includes: The upper-level capacity optimization model is solved using a non-dominated sorting genetic algorithm to obtain a non-dominated solution set that includes multiple capacity configuration schemes. For each capacity configuration scheme in the non-dominated solution set, the particle swarm optimization algorithm is used to solve the lower-level scheduling optimization model to obtain the optimal scheduling scheme and operating cost corresponding to the capacity configuration scheme. The operating cost is fed back to the upper-level capacity optimization model to determine the system operating cost in the annualized total cost.

9. The method for optimizing the capacity configuration of a large-scale wind-solar-storage-hydrogen production dual-layer system according to claim 1, characterized in that, In step S6, the multi-attribute decision-making method is an ideal solution method. It selects the target capacity configuration scheme and the corresponding optimal scheduling scheme from the non-dominated solution set using the multi-attribute decision-making method, including: Calculate the Euclidean distance from each capacity configuration scheme to the positive ideal solution and the negative ideal solution. Calculate the relative proximity of each capacity configuration scheme based on the Euclidean distance. Determine the capacity configuration scheme with the highest relative proximity and its corresponding preferred scheduling scheme as the target capacity configuration scheme and its corresponding preferred scheduling scheme.

10. A large-scale wind-solar-storage-hydrogen production dual-layer capacity configuration optimization device, characterized in that, include: The scenario construction module is used to construct a large-scale wind-solar-storage-hydrogen production system model based on pre-acquired wind and solar power output scenarios and hydrogen load scenarios. The large-scale wind-solar-storage-hydrogen production system model includes a power subsystem, a hydrogen subsystem, and an electrolyzer array. A two-layer model construction module is used to construct a two-layer capacity configuration optimization model, which includes an upper-layer capacity optimization model and a lower-layer scheduling optimization model. The upper-level optimization module is used to optimize the annualized total cost, the curtailment rate, and the energy supply reliability, and sets equipment capacity configuration constraints to solve the capacity configuration scheme of each energy device in the power subsystem, hydrogen subsystem, and electrolyzer array. The lower-level optimization module is used to minimize the overall volatility of the electrolytic cell array as the optimization objective and to set system operation constraints to solve the optimal scheduling scheme under a given capacity configuration scheme; wherein, the overall volatility index is obtained by weighted calculation of multiple factors characterizing the start-up and shutdown characteristics and the stability of operating condition transitions of the electrolytic cell array. The two-layer solution module is used to solve the upper-layer capacity optimization model using a first optimization algorithm to obtain a non-dominated solution set including multiple capacity configuration schemes. For each capacity configuration scheme in the non-dominated solution set, the second optimization algorithm is used to solve the lower-layer scheduling optimization model to obtain the corresponding optimal scheduling scheme and operating cost. The decision output module is used to select the target capacity configuration scheme and the corresponding optimal scheduling scheme from the non-dominated solution set using a multi-attribute decision method based on all capacity configuration schemes and the corresponding optimal scheduling schemes.

Images