Multi-target random capacity configuration method for wind-solar complementary multi-tank collaborative hydrogen production system

Abstract

The invention provides a multi-target random capacity configuration method for a wind-solar complementary multi-tank cooperative hydrogen production system, and relates to the technical field of renewable energy hydrogen production. According to the method, under the uncertain condition, multi-dimensional key system indexes are considered, the single-cell power and number of PEM electrolytic cells and ALK electrolytic cells are finely matched, and a reasonable collaborative operation strategy is designed. On the basis of capacity configuration scheme design and multi-tank collaborative operation strategy optimization of the wind-solar complementary electrolytic hydrogen production system, a multi-target stochastic programming model considering wind-solar fluctuation output is developed, and collaborative optimization of continuous variables of rated capacity and operation power of each device and discrete variables of types, number and start-stop strategies of electrolytic tanks is realized. Aiming at the current situation that the current capacity configuration research mostly stays at the overall capacity ratio level and is lack of discussion on coupling optimization of the capacity, the number and the operation strategy of a single tank, the economic-energy efficiency-hydrogen production triple indexes of the system are effectively improved, and the global optimization of the capacity configuration of the all-off-grid wind-solar electrolytic hydrogen production system is realized.

Description

Technical Field

[0001] This invention relates to the field of renewable energy hydrogen production technology, and in particular to a multi-objective random capacity configuration method for a wind-solar hybrid multi-cell coordinated hydrogen production system. Background Technology

[0002] Currently, the installed capacity of renewable energy sources such as wind power and photovoltaics is experiencing explosive growth, but the resulting problems of wind and solar curtailment are becoming increasingly prominent. The structural contradiction between the volatility of wind and solar power output and the demand for renewable energy power consumption has become a key bottleneck restricting the large-scale and efficient utilization of renewable energy. Hydrogen energy, as a clean energy carrier and an important chemical raw material, can play a role in long-term, large-scale energy storage in new power systems, especially in smoothing wind and solar power fluctuations and achieving cross-seasonal energy dispatch. By constructing an integrated energy system that complements electricity and hydrogen, and coordinating electrochemical energy storage with hydrogen energy storage, the capacity for renewable energy consumption can be effectively improved, and energy balance can be achieved across multiple time scales.

[0003] However, fully off-grid wind-solar coupled electrolysis hydrogen production systems still face the dual challenges of low energy conversion efficiency and high total system cost. On the one hand, the randomness and intermittency of wind and solar resources place extremely high demands on the dynamic response capability, load regulation range, and start-up and shutdown frequency of electrolyzers. On the other hand, existing capacity configuration methods mostly focus on optimizing the overall capacity ratio, lacking in-depth analysis of the coupling relationship between single-cell capacity, quantity, and operating strategies. Among the current mainstream electrolyzer technologies, alkaline (ALK) electrolyzers have the advantage of lower cost, but their hydrogen production efficiency is relatively low, and their start-up speed is slow, making it difficult to adapt to rapid fluctuations in wind and solar power output. Proton exchange membrane (PEM) electrolyzers, while possessing high efficiency, rapid start-up and shutdown, and wide load regulation capabilities, are more suitable for handling fluctuating loads, but their unit capacity investment cost is approximately two to three times that of ALK electrolyzers. Therefore, the mixed use of ALK and PEM electrolyzers, complementing each other's advantages, has become an important way to improve the system's economy and operational flexibility.

[0004] Unfortunately, most existing studies only optimize the overall capacity configuration, failing to fully consider the synergistic effects between individual cell specifications, quantity configuration, and operating strategies. While increasing the capacity of individual cells can reduce unit investment costs, it exacerbates efficiency losses and reduces dynamic response capabilities under partial load, leading to a deterioration in overall performance during actual fluctuating operation. Therefore, under conditions of uncertain wind and solar power output, a systematic solution is still lacking for how to precisely configure the power and quantity of PEM and ALK electrolyzers and design matching synergistic operating strategies to achieve multi-objective optimization of system economy, energy efficiency, and hydrogen production.

[0005] In summary, there is an urgent need to propose a multi-objective stochastic programming method that can perform coupled decision-making on single-tank capacity-quantity and start-up / shutdown and load allocation strategies, so as to achieve global optimization of capacity configuration and operation strategy of electrolytic hydrogen production system under multiple uncertainties. Summary of the Invention

[0006] To address this issue, this invention provides a multi-objective random capacity configuration method for a wind-solar hybrid multi-cell synergistic hydrogen production system. This method solves the problem in the prior art where the uncertainty of wind and solar power output and the lack of coordinated optimization of the capacity, quantity, and operation strategy of individual electrolyzers lead to an inability to simultaneously achieve system economy, energy efficiency, and hydrogen production.

[0007] To address the aforementioned technical problems, embodiments of the present invention provide a multi-objective stochastic capacity configuration method for a wind-solar hybrid multi-cell coordinated hydrogen production system, the method comprising: S1: Based on the national wind and meteorological database, obtain historical data on wind speed and light intensity at the target location to generate and reduce scenes, and obtain a set of typical wind and meteorological scenes with probability distribution; S2: Construct a multi-cell collaborative hydrogen production system model that includes wind turbines, photovoltaics, batteries, hydrogen storage tanks, alkaline electrolyzers, and proton exchange membrane electrolyzers, and set system operation constraints. The system model includes a collaborative operation strategy for alkaline electrolyzers and proton exchange membrane electrolyzers, which is used to dynamically adjust the start-up and shutdown of electrolyzers and load distribution according to the fluctuations in wind and solar power output. S3: Based on the system model and the set of typical wind and solar scenes, calculate the total annual cost of the system, the system energy efficiency, and the average hydrogen production. S4: Based on the calculation results and the probability distribution of the typical wind and solar scene set, establish a multi-objective stochastic programming model with the optimization objectives of minimizing the total annual cost of the system, maximizing the system energy efficiency, and maximizing the average hydrogen production. The model collaboratively optimizes the continuous variables of rated capacity and operating power of each device and the discrete variables of electrolyzer type, quantity, and start-up / load allocation strategy. S5: The multi-objective stochastic programming model is transformed into a deterministic multi-objective optimization problem using scenario analysis. The deterministic optimization problem is solved using a multi-objective optimization algorithm to obtain the Pareto optimal solution set. Finally, the system capacity configuration and operation strategy scheme are determined from the Pareto optimal solution set using a multi-attribute decision method.

[0008] Preferably, in step S1, the scene generation and reduction includes: preprocessing and extracting features from historical landscape data, clustering landscape features using a clustering method, adaptively selecting feature extraction method, clustering method and its hyperparameters using a Bayesian optimization method, obtaining typical landscape scenes based on the nearest center distance retention strategy within clusters and calculating their probability distribution.

[0009] Preferably, in step S2, the system operation constraints include: upper and lower limits of equipment operation, power balance constraints, hydrogen flow balance constraints, and dynamic energy conservation constraints of energy storage and hydrogen storage equipment.

[0010] Preferably, in step S2, the cooperative operation strategy specifically includes: Alkaline electrolyzers should be given priority to bear the basic load, while proton exchange membrane electrolyzers should bear the peak load. Net power of wind and solar power after energy storage regulation First, the upper limit of the alkaline electrolytic cell installation. contrast; like Not exceeding And not less than the starting power threshold of any alkaline electrolyzer. Then all the power will be allocated to the alkaline electrolytic cell; like Below the starting power threshold of alkaline electrolyzer Then determine whether the starting power threshold of the proton exchange membrane electrolyzer is met. : If lower Then the electricity will be wasted; If higher Then start the proton exchange membrane electrolyzer and operate at rated power. Running, remaining power Handle as follows: If the remaining power Exceeding the maximum overload power of the proton exchange membrane electrolyzer In this case, an additional proton exchange membrane electrolyzer will be added. If the maximum overload power of the proton exchange membrane electrolyzer is not exceeded. In this case, the existing proton exchange membrane electrolyzer will operate under overload and share the remaining power. like If the installed capacity of alkaline electrolyzers exceeds the upper limit, all alkaline electrolyzers will operate at their rated power. During operation, the excess power is distributed sequentially to the proton exchange membrane electrolyzer or abandoned according to the above logic.

[0011] Preferably, in step S4, the discrete variables include the selection of the type and quantity of ALK and PEM electrolyzers, as well as the selection of start-up and shutdown and load distribution strategies; the continuous variables include the rated capacity and real-time operating power of the wind turbine, photovoltaic system, ALK electrolyzer, PEM electrolyzer, battery, and hydrogen storage tank.

[0012] Preferably, in step S4, when establishing the multi-objective stochastic programming model, an economic and performance index regression model based on the functional relationship between the unit cost of the electrolyzer and the hydrogen production efficiency as a function of scale is also introduced to provide parameter support for the collaborative optimization decision of the discrete and continuous variables.

[0013] Preferably, in step S5, the multi-objective optimization algorithm includes one of the following: multi-objective genetic algorithm, multi-objective particle swarm optimization algorithm, and gray wolf algorithm.

[0014] Preferably, in step S5, the multi-attribute decision-making method includes one of the following: the approximation of ideal solution ranking method and the linear programming method of multidimensional preference analysis.

[0015] Preferably, in step S3, the total annual cost of the system includes equipment investment cost, operation and maintenance cost, power curtailment penalty cost, and hydrogen sales revenue.

[0016] This invention also provides a multi-objective stochastic capacity configuration system for a wind-solar hybrid multi-tank synergistic hydrogen production system. This system is used to implement the aforementioned multi-objective stochastic capacity configuration method for a wind-solar hybrid multi-tank synergistic hydrogen production system, specifically including: The landscape scene generation unit is used to generate and reduce scenes based on historical data of wind speed and light intensity of the target location obtained from the national landscape meteorological database, so as to obtain a set of typical landscape scenes with probability distribution. The system modeling and constraint setting unit is used to construct a multi-cell collaborative hydrogen production system model that includes wind turbines, photovoltaics, batteries, hydrogen storage tanks, alkaline electrolyzers and proton exchange membrane electrolyzers, and to set system operation constraints. The system model includes a collaborative operation strategy for alkaline electrolyzers and proton exchange membrane electrolyzers, which is used to dynamically adjust the start-up and shutdown of electrolyzers and load distribution according to the fluctuations in wind and solar power output. The system performance evaluation unit is used to calculate the system's total annual cost, system energy efficiency, and average hydrogen production based on the system model and the typical wind and solar scene set. An optimization modeling unit is used to establish a multi-objective stochastic programming model based on the calculation results and the probability distribution of the typical wind and solar scene set, with the optimization objectives of minimizing the total annual cost of the system, maximizing the system energy efficiency, and maximizing the average hydrogen production. The model collaboratively optimizes the continuous variables of rated capacity and operating power of each device and the discrete variables of electrolyzer type, number, and start-up / load allocation strategy. The decision application unit is used to transform the multi-objective stochastic programming model into a deterministic multi-objective optimization problem using scenario analysis, solve the deterministic optimization problem using a multi-objective optimization algorithm to obtain a Pareto optimal solution set, and use a multi-attribute decision method to determine the final system capacity configuration and operation strategy scheme from the Pareto optimal solution set.

[0017] As can be seen from the above technical solutions, this invention application has the following beneficial effects: (1) This invention proposes an adaptive wind and solar scene analysis method based on Bayesian optimization, which can autonomously complete the optimization of feature extraction methods, scene clustering methods and their hyperparameters. By acquiring historical data from a national wind and solar meteorological database, data preprocessing, low-dimensional feature extraction, cluster analysis and scene reduction are performed to efficiently generate a set of typical wind and solar scenes with probability distribution. This comprehensively and accurately portrays the uncertainty of wind and solar power output, providing reliable data support for subsequent system capacity configuration and operation strategy optimization, and avoiding optimization deviations caused by inaccurate scene analysis.

[0018] (2) This invention constructs a model of a wind-solar complementary multi-cell synergistic electrolysis hydrogen production system, which includes mathematical models of core equipment such as wind turbines, photovoltaics, PEM electrolyzers, ALK electrolyzers, batteries, and hydrogen storage tanks, and clarifies the operating constraints, power conservation constraints, and hydrogen flow conservation constraints of each equipment. At the same time, it innovatively proposes a synergistic operation strategy in which the ALK electrolyzer undertakes the base load and the PEM electrolyzer is responsible for the peak load. The start-up and shutdown of the electrolyzers and the load distribution are dynamically adjusted according to the fluctuations in wind and solar power output, giving full play to the advantages of the low cost of the ALK electrolyzer and the flexible start-up and shutdown and strong peak-shaving capability of the PEM electrolyzer, effectively reducing the impact of wind and solar fluctuations on the hydrogen production system, and ensuring a stable increase in average hydrogen production while improving the system's energy utilization efficiency.

[0019] (3) To address the issue that existing research often focuses on overall capacity allocation and lacks coupled optimization of single-cell capacity-quantity-operation strategy, a multi-objective stochastic programming model is established. Within the same optimization framework, discrete variables such as electrolyzer type, quantity, and start-up / load allocation strategies, as well as continuous variables such as equipment rated capacity and real-time operating power, are solved collaboratively. A data model based on the unit cost of electrolyzers and the variation of hydrogen production efficiency with scale is introduced to provide parameter support for hybrid decision-making. By using the scenario method, the stochastic programming problem is transformed into a deterministic problem. Combining multi-objective optimization algorithms and multi-attribute decision-making methods, the collaborative optimization of the three objectives of total annual cost, energy efficiency, and hydrogen production is finally achieved, resulting in the global optimum of capacity configuration and operation strategy for the off-grid wind-solar electrolysis hydrogen production system. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly described below. Referring to the drawings will make the features and advantages of the present invention clearer. The drawings are illustrative and should not be construed as limiting the present invention in any way. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1This is a flowchart of a multi-objective random capacity configuration method for a wind-solar hybrid multi-cell synergistic hydrogen production system provided by the present invention; Figure 2 This is a schematic diagram of the structure of the wind-solar hybrid multi-cell synergistic hydrogen production system in this invention; Figure 3 This is a flowchart of the multi-objective stochastic programming solution strategy in this invention; Figure 4 This is a diagram of the PEM-ALK synergistic hydrogen production operation strategy in this invention; Figure 5 This is a block diagram of a multi-objective random capacity configuration system for a wind-solar hybrid multi-cell coordinated hydrogen production system provided by the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Example 1: To address the issue of multi-objective random capacity configuration in existing wind-solar hybrid multi-tank co-production hydrogen systems, such as... Figure 1 As shown, this invention proposes a multi-objective stochastic capacity configuration method for a wind-solar hybrid multi-cell synergistic hydrogen production system, which includes: S1: Based on the national wind and meteorological database, obtain historical data on wind speed and light intensity at the target location to generate and reduce scenes, and obtain a set of typical wind and meteorological scenes with probability distribution; S2: Construct a multi-cell collaborative hydrogen production system model that includes wind turbines, photovoltaics, batteries, hydrogen storage tanks, alkaline electrolyzers, and proton exchange membrane electrolyzers, and set system operation constraints. The system model includes a collaborative operation strategy for alkaline electrolyzers and proton exchange membrane electrolyzers, which is used to dynamically adjust the start-up and shutdown of electrolyzers and load distribution according to the fluctuations in wind and solar power output. S3: Based on the system model and a set of typical wind and solar scenarios, calculate the system's total annual cost, system energy efficiency, and average hydrogen production. S4: Based on the calculation results and the probability distribution of typical wind and solar scene sets, a multi-objective stochastic programming model is established with the optimization objectives of minimizing the total annual cost of the system, maximizing the system energy efficiency, and maximizing the average hydrogen production. The model co-optimizes the continuous variables of rated capacity and operating power of each device, as well as the discrete variables of electrolyzer type, quantity, and start-up / load allocation strategy. S5: The multi-objective stochastic programming model is transformed into a deterministic multi-objective optimization problem using scenario analysis. The deterministic optimization problem is solved using a multi-objective optimization algorithm to obtain the Pareto optimal solution set. Finally, the system capacity configuration and operation strategy scheme are determined from the Pareto optimal solution set using a multi-attribute decision method.

[0023] As can be seen from the above technical solution, this invention proposes a multi-objective stochastic capacity configuration method for a wind-solar hybrid multi-tank synergistic hydrogen production system. It generates a set of typical wind and solar scenarios with probability distributions based on historical wind and solar data, providing accurate uncertainty data support for subsequent optimization; constructs a multi-tank synergistic hydrogen production system model containing core equipment and operational constraints, establishing the basic framework and boundary conditions for system optimization; calculates the system's annual total cost, energy efficiency, and average hydrogen production, providing quantitative performance evaluation basis for the optimization model; establishes a multi-objective stochastic programming model for synergistic optimization of discrete and continuous variables, achieving coupled optimization of the system's economic, energy efficiency, and hydrogen production objectives; transforms the stochastic model into a deterministic problem and solves it to obtain the optimal configuration scheme, efficiently outputting feasible decisions that take into account multiple objectives, ultimately achieving global optimization of the capacity configuration and operation strategy of the fully off-grid wind-solar electrolysis hydrogen production system.

[0024] The wind-solar hybrid multi-cell synergistic hydrogen production system targeted by this invention is as follows: Figure 2 As shown, it consists of the following five modules: (1) Renewable energy generation module, including photovoltaic and wind turbine generator sets; (2) Energy storage module, which uses batteries to realize electrochemical energy storage; (3) Multi-cell hydrogen production module, which uses PEM and ALK multi-cell combination for electrolytic hydrogen production; (4) Hydrogen storage module, which uses hydrogen storage tanks to store the prepared hydrogen; (5) Hydrogen consumption module, which includes the hydrogen consumption needs of downstream plants and the sale of surplus hydrogen.

[0025] The wind-solar hybrid multi-cell coordinated hydrogen production system proposed in this invention aims to reduce system costs and improve system energy utilization efficiency by coordinating the capacity configuration and operation strategy of internal equipment to meet the hydrogen demand of downstream plants. It utilizes energy storage modules to smooth out power output fluctuations from wind and solar power, employs a combined ALK electrolyzer and PEM electrolyzer to fully leverage the advantages of low cost of the ALK electrolyzer and high efficiency and flexible start-stop of the PEM electrolyzer, and uses hydrogen storage modules to store surplus hydrogen for further profit from hydrogen sales.

[0026] The following is combined Figure 3 The technical solution of the present invention will be described in detail below.

[0027] In step S1, historical data on wind speed and light intensity at the target location are collected from a national wind and solar meteorological database to generate and reduce scenes, resulting in a set of typical wind and solar scenes with probability distributions. Scene generation and reduction include: preprocessing and extracting features from historical wind and solar data; clustering wind and solar features using a clustering method; adaptively selecting the feature extraction method, clustering method, and hyperparameters using a Bayesian optimization method; and obtaining typical wind and solar scenes based on a nearest-center-distance retention strategy within clusters and calculating their probability distributions.

[0028] Specifically, step S1 includes the following steps: Step S11: Preprocessing of historical wind and solar data: Based on the national wind and solar meteorological database, historical data of wind speed and light intensity of the target location are obtained. After reading the original wind and solar sequences and unifying the time base, the data of many years are sliced ​​by day using a 24×1h sliding window. Then, missing measurements are filled in using linear interpolation, and outliers are removed using the interquartile range (IQR) method. Finally, the data is normalized using the Min-Max method to provide clean and comparable standardized input for subsequent modeling.

[0029] Step S12: Scenery feature extraction: Use linear dimensionality reduction method PCA or nonlinear UMAP dimensionality reduction method to extract low-dimensional features from the normalized high-dimensional historical scenery data.

[0030] Step S13: Scenery Clustering: Clustering methods such as K-means, K-medoids, or spectral clustering are used to perform cluster analysis on the low-dimensional features of the scenery, and the silhouette coefficients are calculated to evaluate the quality of the clustering results. The method for calculating the silhouette coefficients is as follows: ; In the formula, , , Sample points Profile coefficient, The average distance to other points in the same cluster and the average distance to the nearest different cluster; The overall profile coefficient has a range of values ​​of [value range missing]. When the silhouette coefficient is greater than 0.5, the clustering effect is considered acceptable, and when it is greater than 0.7, the clustering effect is considered excellent.

[0031] Step S14: Adaptively adjust the dimensionality reduction, clustering methods, and hyperparameters based on Bayesian optimization to optimize clustering performance. Optimized variables include, but are not limited to: feature extraction methods PCA or UMAP, scene clustering methods K-means, K-medoids, or spectral clustering; and their corresponding hyperparameters such as dimensionality reduction dimension n_components, number of nearest neighbors n_neighbors, minimum distance min_dist, and number of clusters K. The basic process of Bayesian optimization is as follows: First, using cross-validation error as the objective function, construct a Gaussian process prior for the high-dimensional hyperparameter space; during iteration, select the next evaluation point for true evaluation based on the maximum expectation improvement or upper confidence bound acquisition function, then update the posterior distribution, and converge to the optimal or near-optimal solution after several rounds.

[0032] Step S15: Scene Reduction: Based on the nearest center distance retention strategy within a cluster, the scenery scene closest to the cluster center is retained as the typical scenery scene of that cluster; finally, K typical scenery scenes and their probability distributions are obtained. A typical scenario The probability can be calculated based on empirical frequencies, i.e., the first... The ratio of the number of samples within a cluster to the total number of samples in a class: ; In the formula, For the first The ratio of the number of samples within a cluster to the total number of samples in a class. For the first The number of samples within a cluster of a class. This represents the total number of samples.

[0033] In step S2, a multi-cell collaborative hydrogen production system model is constructed, including a wind turbine, photovoltaic system, battery, hydrogen storage tank, alkaline electrolyzer, and proton exchange membrane electrolyzer. System operation constraints are set, including: upper and lower limits of equipment operation, power balance constraints, hydrogen flow balance constraints, and dynamic energy conservation constraints of energy storage and hydrogen storage equipment.

[0034] The power output models for wind turbines and photovoltaic systems are as follows: ; ; In the formula, These are the photovoltaic rated capacity, real-time output, rated irradiance, and real-time irradiance, respectively. , , , , and These are the fan's rated capacity, real-time output, real-time wind speed, cut-in wind speed, cut-out wind speed, and rated wind speed, respectively.

[0035] The battery energy storage model is as follows: ; In the formula, , , , and These are the battery's real-time energy storage capacity, charging efficiency, discharging efficiency, real-time charging power, and discharging power, respectively. This is a binary variable representing the charging and discharging state of a battery, indicating that the battery cannot be charged and discharged simultaneously.

[0036] In addition, the battery should meet the following constraints during operation: real-time capacity is less than rated capacity, charging and discharging processes do not occur simultaneously, charging and discharging power is less than the threshold, and the initial and final energy storage states are equal. .

[0037] The hydrogen production models for alkaline (ALK) electrolyzers and proton exchange membrane (PEM) electrolyzers are as follows: ; In the formula, and For the first The first PEM electrolytic cell and the first Real-time operating power of each ALK electrolyzer. , , and The hydrogen production rates and efficiencies of PEM and ALK electrolyzers are presented respectively. Furthermore, considering the differences in start-up and shutdown procedures between PEM and ALK electrolyzers, a start-up / shutdown penalty term is introduced. This indicates the percentage of startup time within the runtime period. express The number of newly added electrolytic cells during the period, This indicates the average hydrogen production. It is the high calorific value of hydrogen, that is, the total heat released when the water vapor produced by the complete combustion of a unit of hydrogen gas condenses into liquid water.

[0038] Furthermore, during operation, both PEM and ALK electrolyzers must meet power operation limits for a single cell, and the PEM electrolyzer employs rolling optimization, allowing only a maximum of half of the cells to be overloaded. .

[0039] Furthermore, this invention designs a collaborative operation strategy for ALK electrolyzers and PEM electrolyzers. This strategy is configured such that the ALK electrolyzers preferentially bear the base load, while the PEM electrolyzers bear the peak-shaving load; and the start-up and shutdown of the electrolyzers and load distribution are dynamically adjusted according to fluctuations in wind and solar power output. Specifically, as... Figure 4 As shown, the net power of wind and solar power after energy storage regulation First, the upper limit of ALK electrolytic cell installation. In comparison, if the limits are not exceeded and the starting power threshold of any ALK electrolyzer is met... (In this implementation) 20% If the power is insufficient to start the ALK electrolyzer, then all power is allocated to the ALK electrolyzer; if it is insufficient to start the ALK electrolyzer, then check if the power threshold for starting the PEM electrolyzer is met. (In this implementation) 5% If the power consumption is still below the PEM electrolyzer startup threshold, power will be withheld; if startup is possible, the PEM electrolyzer will be powered primarily at its rated power. Running, the remaining power at this time If the maximum overload power is exceeded (In this implementation) 25% If the maximum overload power is not exceeded, then an additional PEM electrolytic cell will be added; if the maximum overload power is not exceeded, then all existing PEM electrolytic cells will operate under overload and share the additional power; and when the net power... When the installed capacity of ALK electrolyzers exceeds the upper limit, all ALK electrolyzers will operate at rated power. The system operates, and then the over-generated power is judged in the same way to determine whether it is sufficient to start the PEM electrolytic cell, whether the number of cells needs to be increased, or whether overload operation is allowed. The portion that still cannot be consumed is finally discarded.

[0040] The hydrogen storage tank model is as follows: ; In the formula, , , , and These are the real-time capacity, hydrogen storage efficiency, hydrogen release efficiency, and real-time hydrogen charging / discharging rate of the hydrogen storage tank, respectively. This is a binary variable representing the hydrogen storage and release states of a hydrogen storage tank, indicating that the hydrogen storage tank cannot store or release hydrogen simultaneously.

[0041] In addition, hydrogen storage tanks should meet the following constraints during operation: real-time capacity is less than rated capacity, hydrogen storage and release processes do not occur simultaneously, the hydrogen storage and release rates are less than a threshold, and the initial and final hydrogen storage states are equal. .

[0042] Furthermore, the hydrogen production system should also comply with the power conservation constraint and the hydrogen flow rate conservation constraint as follows:

[0043] In the formula, This refers to the power that has been abandoned. For the amount of hydrogen used, Hydrogen sales will only be conducted at the end of the day to determine the volume of hydrogen sold.

[0044] In step S3, based on the above system model and typical wind and solar scene set, the total annual cost of the system, the system energy efficiency and the average hydrogen production are calculated to evaluate the economic and technical performance of the system.

[0045] Specifically, the system's total annual cost The costs include equipment investment, operation and maintenance, curtailment penalty costs, and hydrogen sales revenue, calculated as follows: ; In the formula, and These are the equipment investment cost and the operation and maintenance cost, respectively. and These are the equipment's rated capacity and the equipment's real-time operating power, respectively. For the number of days, and These represent the investment cost and operation and maintenance cost per unit capacity, respectively. The discount rate is... For equipment lifespan; The penalty cost per unit of electricity wasted. This refers to the price of hydrogen sold.

[0046] System energy efficiency The calculation method is as follows: ; In the formula, It is the high calorific value of hydrogen, that is, the total heat released when the water vapor produced by the complete combustion of a unit of hydrogen gas condenses into liquid water.

[0047] Average hydrogen production of the system The calculation method is as follows: .

[0048] In step S4, based on the calculation results and the probability distribution of typical wind and solar scenarios, a multi-objective stochastic programming model is established with the optimization objectives of minimizing the total annual system cost, maximizing the system energy efficiency, and maximizing the average hydrogen production. This model collaboratively optimizes continuous and discrete variables in system design and operation. Discrete variables include the selection of ALK and PEM electrolyzer types, quantity configurations, and start-up / shutdown and load allocation strategies; continuous variables include the rated power and real-time operating power of wind turbines, photovoltaic systems, ALK electrolyzers, PEM electrolyzers, batteries, and hydrogen storage tanks.

[0049] Furthermore, the multi-objective stochastic programming model is expressed as: ; Furthermore, the decision variables include the rated capacity and operating variables of the equipment, where the rated capacity includes the rated power of wind turbines, photovoltaic systems, PEM single-cell systems, and ALK single-cell systems. , , , The number of PEM and ALK electrolyzers and Rated capacity of storage batteries and hydrogen storage tanks and The manipulated variables include the real-time output of photovoltaic and wind turbines. The number of electrolytic cells started and stopped in real time / Load Distribution Strategy Battery charging and discharging power , abandoned power and the hydrogen storage and release rate of the hydrogen storage tank. , The constraints include upper and lower bound constraints on the capacity and continuous values ​​of the operating variables, discrete variable constraints on the types and quantities of electrolytic cells, constraints on equipment operation processes, and energy conservation constraints.

[0050] Furthermore, the aforementioned multi-objective stochastic programming model is based on a discrete-continuous hybrid decision-making approach. Within the same optimization framework, it collaboratively solves for discrete variables such as the type and quantity of electrolyzers and start-up / load allocation strategies, as well as continuous variables such as the real-time input power of each device, to obtain the globally optimal configuration and operation scheme. Given that the unit cost of PEM and ALK electrolyzer stacks decreases non-linearly with increasing installed capacity, and their electrochemical polarization losses increase synchronously, leading to a decrease in hydrogen production efficiency, under the premise of the same total rated capacity, a small number of large-capacity single-unit stacks, while having lower equipment costs, have slightly lower energy conversion efficiency and limited dynamic response. Conversely, multiple sets of small-capacity single-unit stacks, while having higher equipment costs, offer excellent energy conversion efficiency and possess rapid start-up and shutdown capabilities and flexible peak-shaving capabilities. This invention constructs regression data models based on market and literature data to analyze the functional relationship between the unit cost of a single stack and hydrogen production efficiency with scale, respectively, to accurately estimate its economic efficiency and performance indicators within any given scale range, thereby providing reliable data support for the aforementioned discrete-continuous hybrid decision-making approach.

[0051] In step S5, the multi-objective stochastic programming model is transformed into a deterministic multi-objective optimization problem using scenario analysis. The deterministic optimization problem is then solved using a multi-objective optimization algorithm to obtain the Pareto optimal solution set. Finally, the system capacity configuration and operation strategy scheme are determined from the Pareto optimal solution set using a multi-attribute decision method.

[0052] Specifically, using the typical landscape scenes and their probability distribution generated in step S1 as input, the uncertainty of landscape is transformed into the expected value under finite discrete scenes, thereby mapping the original stochastic optimization problem into a solvable deterministic multi-objective optimization problem.

[0053] Furthermore, multi-objective optimization methods can employ multi-objective particle swarm optimization, multi-objective genetic algorithm, gray wolf algorithm, and other multi-objective optimization algorithms to solve the above-mentioned multi-objective problems. Taking the NSGA-II algorithm as an example, the Pareto optimal solution set can be obtained efficiently and stably through the processes of population initialization, offspring generation, non-dominated sorting, crowding distance sorting, and elites.

[0054] Furthermore, multi-objective decision-making methods can employ TOPSIS, LINMAP, and other approaches. Based on scenario preferences, reasonable weights are set to achieve the final decision on feasible configurations. Taking TOPSIS as an example, the original values ​​of each solution under each indicator are first normalized and weighted. Then, the optimal combination of all indicators is identified as the positive ideal solution, and the worst combination as the negative ideal solution. The Euclidean distance between each solution and these two solutions is calculated, and finally, the solutions are ranked by proximity, with the solution having the highest proximity being selected as the optimal decision.

[0055] Example 2: Figure 5As shown, this invention provides a multi-objective stochastic capacity configuration system for a wind-solar hybrid multi-tank synergistic hydrogen production system. This system is used to implement the multi-objective stochastic capacity configuration method for the wind-solar hybrid multi-tank synergistic hydrogen production system described in Embodiment 1 above, specifically including: The landscape scene generation unit 100 is used to obtain historical data on wind speed and light intensity of the target location from the national landscape meteorological database, generate and reduce the scene to obtain a set of typical landscape scenes with probability distribution. The system modeling and constraint setting unit 200 is used to construct a multi-cell collaborative hydrogen production system model that includes wind turbines, photovoltaics, batteries, hydrogen storage tanks, alkaline electrolyzers and proton exchange membrane electrolyzers, and to set system operation constraints. The system model includes a collaborative operation strategy for alkaline electrolyzers and proton exchange membrane electrolyzers, which is used to dynamically adjust the start-up and shutdown of electrolyzers and load distribution according to the fluctuations in wind and solar power output. The system performance evaluation unit 300 is used to calculate the system's total annual cost, system energy efficiency, and average hydrogen production based on the system model and a set of typical wind and solar scenarios. The optimization modeling unit 400 is used to establish a multi-objective stochastic programming model based on the calculation results and the probability distribution of typical wind and solar scene sets. The model aims to minimize the total annual cost of the system, maximize the energy efficiency of the system, and maximize the average hydrogen production. The model collaboratively optimizes the continuous variables of rated capacity and operating power of each device, as well as the discrete variables of electrolyzer type, number, and start-up / load allocation strategy. The decision application unit 500 is used to transform a multi-objective stochastic programming model into a deterministic multi-objective optimization problem using scenario analysis, solve the deterministic optimization problem using a multi-objective optimization algorithm to obtain the Pareto optimal solution set, and use a multi-attribute decision method to determine the final system capacity configuration and operation strategy scheme from the Pareto optimal solution set.

[0056] This embodiment presents a multi-objective random capacity configuration system for a wind-solar hybrid multi-tank co-production hydrogen system, used to implement the aforementioned multi-objective random capacity configuration method for a wind-solar hybrid multi-tank co-production hydrogen system. Therefore, the specific implementation of this system can be found in the embodiments section of the aforementioned multi-objective random capacity configuration method for a wind-solar hybrid multi-tank co-production hydrogen system. For example, the wind-solar scene generation unit 100, system modeling and constraint setting unit 200, system performance evaluation unit 300, optimization modeling unit 400, and decision application unit 500 are respectively used to implement steps S1, S2, S3, S4, and S5 in the aforementioned multi-objective random capacity configuration method for a wind-solar hybrid multi-tank co-production hydrogen system. Therefore, its specific implementation can be referred to the descriptions of the corresponding embodiments. To avoid redundancy, further details are omitted here.

[0057] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied 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.

[0058] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0059] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0060] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for multi-objective random capacity configuration of a wind-solar complementary multi-tank collaborative hydrogen production system, characterized in that, The method comprises the following steps: S1: According to the national wind and light weather database, the historical data of wind speed and light intensity of the target place are collected to generate and reduce the scene, and a typical wind and light scene set with probability distribution is obtained; S2: A multi-tank collaborative hydrogen production system model including wind turbines, photovoltaic, batteries, hydrogen storage tanks, and alkaline electrolyzers and proton exchange membrane electrolyzers is constructed, and system operation constraints are set, wherein the system model includes a collaborative operation strategy of alkaline electrolyzers and proton exchange membrane electrolyzers, which is used to dynamically adjust the start-stop and load distribution of electrolyzers according to wind and light output fluctuations; S3: Based on the system model and the typical wind and light scene set, the total annual cost of the system, the energy efficiency of the system, and the average hydrogen production are calculated; S4: Based on the calculation results and the probability distribution of the typical wind and light scene set, a multi-objective random programming model is established, with the optimization objectives of minimizing the total annual cost of the system, maximizing the energy efficiency of the system, and maximizing the average hydrogen production. The model collaboratively optimizes the continuous variables of the rated capacity and operating power of each device and the discrete variables of the type, number, start-stop / load distribution strategy of the electrolyzer; S5: The scenario analysis method is used to convert the multi-objective random programming model into a deterministic multi-objective optimization problem, and a multi-objective optimization algorithm is used to solve the deterministic optimization problem to obtain a Pareto optimal solution set. A multi-attribute decision-making method is used to determine the final system capacity configuration and operation strategy scheme from the Pareto optimal solution set.

2. The multi-objective stochastic capacity configuration method of the wind-solar complementary multi-tank coordinated hydrogen production system according to claim 1, characterized in that, In step S1, the scene generation and reduction includes: preprocessing and feature extraction of wind and light historical data, clustering wind and light features using clustering methods, and adaptively selecting feature extraction methods, clustering methods and their hyperparameters using Bayesian optimization method. The typical wind and light scenes are obtained based on the nearest center distance retention strategy and the probability distribution is calculated. 3.The method according to claim 1, wherein, In step S2, the system operation constraints include: device operation upper and lower limit constraints, power balance constraints, hydrogen flow balance constraints, dynamic energy conservation constraints of energy storage and hydrogen storage devices.

4. The method according to claim 1, wherein, In step S2, the collaborative operation strategy specifically includes: The basic load is preferentially borne by the alkaline electrolyzer, and the peak load is borne by the proton exchange membrane electrolyzer. Net power of wind and light after energy storage adjustment Upper limit of alkaline electrolyzer installation Contrast; if not exceeding and not below any basic electrolyzer start-up power threshold then allocate all power to the basic electrolyzer; If the alkaline electrolyzer startup power threshold is exceeded then determine whether the proton exchange membrane electrolyzer startup power threshold is met : If lower than then discard the power; If higher than then start the PEM electrolyser and run at rated power , the remaining power is handled as follows: If the remaining power exceeds the maximum overload power of the proton exchange membrane electrolyzer , then an additional proton exchange membrane electrolyzer is started up If the maximum overload power of the proton exchange membrane electrolyzer is not exceeded then the existing proton exchange membrane electrolyzer is operated under overload and shares the remaining power jointly; If If the installed capacity of alkaline electrolyzer exceeds the upper limit, all alkaline electrolyzers run at rated power and the excess power is distributed to the PEM electrolyzers or abandoned according to the above logic.

5. The method according to claim 1, wherein, In step S4, the discrete variables include the type selection, number configuration, and start-stop and load distribution strategy selection of ALK and PEM electrolyzers; and the continuous variables include the rated capacity and real-time operating power of wind turbines, photovoltaic, ALK electrolyzers, PEM electrolyzers, batteries, and hydrogen storage tanks.

6. The method according to claim 1, wherein, In step S4, when establishing the multi-objective random programming model, an economic and performance index regression model based on the functional relationship between the unit cost and the scale change of the electrolyzer is introduced to provide parameter support for the collaborative optimization decision of the discrete variables and the continuous variables.

7. The method according to claim 1, wherein, In step S5, the multi-objective optimization algorithm includes one of the multi-objective genetic algorithm, the multi-objective particle swarm algorithm, and the grey wolf algorithm. 8.The method of claim 1, wherein, In step S5, the multi-attribute decision-making method includes one of the approximation ideal solution sorting method and the linear programming method of multi-dimensional preference analysis. 9.The method of claim 1, wherein, In step S3, the system annual total cost includes equipment investment cost, operation and maintenance cost, abandoned electricity penalty cost and hydrogen sales revenue.

10. A wind-solar complementary multi-tank collaborative hydrogen production system multi-objective random capacity configuration system, characterized in that, The system is used to realize the multi-objective random capacity configuration method of the wind-solar complementary multi-tank collaborative hydrogen production system according to any one of claims 1 to 9, and specifically comprises: A wind-solar scene generation unit is configured to generate and reduce scenes according to historical wind speed and light intensity data of a target site obtained from a national wind-solar weather database, to obtain a set of typical wind-solar scenes with probability distribution; A system modeling and constraint setting unit is configured to construct a multi-tank collaborative hydrogen production system model including wind turbines, photovoltaic devices, batteries, hydrogen storage tanks, and alkaline electrolysis tanks and proton exchange membrane electrolysis tanks, and set system operation constraints, wherein the system model includes a collaborative operation strategy of the alkaline electrolysis tanks and the proton exchange membrane electrolysis tanks, and is configured to dynamically adjust electrolysis tank start-stop and load distribution according to wind-solar output fluctuations; A system performance evaluation unit is configured to calculate system annual total cost, system energy efficiency and average hydrogen production based on the system model and the set of typical wind-solar scenes; An optimization modeling unit is configured to establish a multi-objective random programming model with the optimization objectives of minimizing system annual total cost, maximizing system energy efficiency and maximizing average hydrogen production based on the calculation results and the probability distribution of the set of typical wind-solar scenes, wherein the model collaboratively optimizes continuous variables of rated capacity and operating power of each device and discrete variables of electrolysis tank type, number, start-stop / load distribution strategy; A decision application unit is configured to convert the multi-objective random programming model into a deterministic multi-objective optimization problem by using a scenario analysis method, to solve the deterministic optimization problem by using a multi-objective optimization algorithm, to obtain a Pareto optimal solution set, and to determine a final system capacity configuration and operation strategy scheme from the Pareto optimal solution set by using a multi-attribute decision-making method.

Images