Many-to-one alkaline electrolytic hydrogen production system separator capacity configuration method and device based on improved multi-objective optimization search algorithm

Abstract

The invention provides a many-for-one alkaline electrolytic hydrogen production system separator capacity design method and device based on an improved multi-objective optimization search algorithm, and belongs to the technical field of hydrogen production system capacity configuration. Comprising the following steps: establishing a thermal-mass coupling dynamic model of an N-in-1 alkaline electrolytic hydrogen production system, including an accumulation dynamic model of oxygen impurity HTO in hydrogen and a thermal inertia model; defining multi-dimensional performance indexes, including starting time, HTO fluctuation index, temperature consistency and low-load operation time; taking the separator capacity as a decision variable, and taking minimization of starting time and HTO fluctuation index and maximization of low-load operation time as optimization objectives to construct a multi-objective optimization model; and solving the multi-objective optimization model by using an improved non-dominated sorting genetic algorithm NSGA-II, obtaining a Pareto optimal solution set of the separator capacity, and obtaining a separator capacity configuration result.

Description

Technical Field

[0001] This invention relates to the field of hydrogen production system capacity configuration technology, and in particular to a method and apparatus for configuring the capacity of a separator in a many-to-one alkaline electrolysis hydrogen production system based on an improved multi-objective optimization search algorithm. Background Technology

[0002] With the large-scale grid integration of renewable energy sources such as wind and solar power, alkaline electrolysis for hydrogen production has become an important technological pathway for consuming green electricity and producing green hydrogen. To improve economic efficiency and compactness, an "N-in-1" structure is often adopted in industry, where multiple electrolyzers share a single-unit (BoP) setup, including separators and heat exchangers. As a key unit for gas-liquid separation, heat storage, and gas buffering, the capacity of the separator directly affects the system's thermal inertia, dynamic response speed, and hydrogen impurity accumulation rate. How to rationally design the separator capacity under the background of renewable energy fluctuations, balancing rapid response with safe and stable operation, remains a challenge in current engineering practice.

[0003] Existing alkaline electrolysis hydrogen production systems face challenges such as large power fluctuations, frequent start-stops, and long periods of low-load operation when integrating renewable energy. As a core auxiliary equipment component (BoP), the capacity design of the separator directly affects the system's dynamic response, thermal inertia, hydrogen impurity accumulation, and operational safety.

[0004] In existing technologies, the capacity configuration of separators in multi-to-one alkaline electrolysis hydrogen production systems relies on estimation methods based on engineering experience or conservative over-design. This approach, lacking precise modeling of the dynamic processes of thermal-mass coupling within the system, fails to establish quantitative relationships between separator capacity and multi-dimensional performance indicators such as start-up time, hydrogen impurity (HTO) fluctuations, temperature uniformity, and low-load operating duration. Consequently, separator capacity design struggles to effectively balance conflicting objectives such as rapid response, safe operation, and low-load adaptability. The direct consequence is that excessively large capacities designed to mitigate risks reduce the system's dynamic response speed and increase investment costs; while capacities determined solely by experience may lead to excessive hydrogen impurities, uneven temperature, or operational instability under renewable energy power fluctuations, failing to simultaneously achieve system flexibility, safety, and economy. Summary of the Invention

[0005] In view of this, the present invention provides a method and apparatus for configuring the capacity of a separator in a multi-to-one alkaline electrolysis hydrogen production system based on an improved multi-objective optimization search algorithm. By accurately modeling the dynamic process of thermal-mass coupling inside the hydrogen production system, the optimization algorithm is used to solve the separator capacity range that meets the multi-objective performance requirements, so that the hydrogen impurities meet the standards, the system operates stably and the temperature is uniform, while taking into account the system's flexibility, safety and economy.

[0006] The technical solution adopted by the embodiments of the present invention to solve its technical problem is as follows:

[0007] The first aspect of this invention provides a method for configuring the capacity of a separator in a many-to-one alkaline electrolysis hydrogen production system based on an improved multi-objective optimization search algorithm, comprising:

[0008] A dynamic model of thermal-mass coupling was established for an N-in-1 alkaline electrolysis hydrogen production system, including a dynamic model of the accumulation of oxygen impurity HTO in hydrogen and a thermal inertia model.

[0009] Define multi-dimensional performance metrics, including startup time, HTO fluctuation index, temperature consistency, and low-load running time;

[0010] Using separator capacity as the decision variable, and minimizing start-up time and HTO fluctuation index, and maximizing low-load operating time as the optimization objectives, a multi-objective optimization model is constructed.

[0011] The improved non-dominated sorting genetic algorithm NSGA-II is used to solve the multi-objective optimization model, obtain the Pareto optimal solution set of the separator capacity, and obtain the separator capacity configuration result.

[0012] Preferably, the establishment of the thermal-mass coupling dynamic model of the N-in-1 alkaline electrolysis hydrogen production system includes:

[0013] The industrial-grade N-in-1 alkaline electrolysis hydrogen production system includes N parallel electrolyzers, a shared gas-liquid separator, a heat exchanger, and a circulating pump;

[0014] HTO Accumulation Dynamic Model: ; ;

[0015] In the formula: This indicates the current flowing through the electrolysis chamber; Indicates the number of electrolysis chambers; It is Faraday efficiency; It is Faraday's constant; , , These represent the molar flow rates of hydrogen impurities, hydrogen, and oxygen at the gas phase outlet of the oxygen-alkali separator, respectively.

[0016] According to the definition of accumulation process : ; ; ; ;

[0017] In the formula, i represents the electrolytic cell number; This indicates the molar flow rate of hydrogen impurities carried by the alkaline solution circulation during the stage from the anode to the cathode. This indicates the molar flow rate of hydrogen impurities transferred by diffusion from the electrolyzer to the liquid phase stage of the separator along with the alkali solution circulation. This represents the molar flow rate of hydrogen impurities transferred through convection mechanisms as they enter the gas phase stage of the separator under the influence of gravity. This indicates the solubility of hydrogen in alkaline solutions; Indicates the flow rate of the alkali solution; Indicates the density of the alkali solution; This represents the diffusion coefficient of hydrogen. This represents the difference in hydrogen concentration between the hydrogen side and the oxygen side; Indicates the thickness of the diaphragm; This indicates the permeability of hydrogen through the membrane; This represents the voltage difference across the half-cell. Indicates the viscosity of the alkaline solution;

[0018] The state-space model of HTO transfer and accumulation process in the N-in-1 system is as follows: ;

[0019] In the formula: , and These represent the number of moles of hydrogen impurities in the anode half-cell, the liquid phase of the separator, and the gas phase of the separator, respectively. , , These represent the molar flow rates of hydrogen impurities transferred from the anode half-cell to the separator liquid phase, from the separator liquid phase to the gas phase, and exiting the gas-liquid separator, respectively, satisfying the following: ;

[0020] In the formula: Indicates the gas-liquid separation time constant; This indicates the volume of the gas phase portion of the gas-liquid separator; It is the ideal gas constant; This refers to the separator outlet temperature.

[0021] The thermal inertia model of the system:

[0022] The thermal balance equation for the electrolytic cell is: ;

[0023] In the formula: This represents the total heat capacity of the electrolytic cell and the alkaline solution within it, i.e. The heat capacity of the alkali solution is calculated as follows: ,in This represents the volume of the alkaline solution in the tank. This is the specific heat capacity of the alkaline solution; and These represent the temperatures of the alkaline solution flowing into and out of the i-th electrolytic cell, respectively. Indicates the heat generated by electrolysis; This indicates that the electrolytic cell dissipates heat to the environment;

[0024] The heat generated by the electrolytic cell consists of the heat of reaction and the ohmic heat generated by the bypass current. ;

[0025] In the formula: It is the thermal neutral voltage; Indicates the number of electrolysis chambers; This refers to the cell voltage; Indicates current; Indicates Faraday efficiency;

[0026] Heat dissipation from the electrolytic cell to the environment Heat dissipation by convection With radiative heat dissipation composition: ;

[0027] In the formula: This represents the equivalent heat dissipation area of ​​the electrolytic cell. The ambient temperature; Boltzmann's constant; Indicates the blackness of the electrolytic cell; The coefficient of heat transfer by natural convection;

[0028] The natural convection heat transfer coefficient is calculated from the size of the electrolytic cell and the temperature difference between it and the environment: ;

[0029] In the formula: The diameter of the electrolytic cell;

[0030] Temperature of alkaline solution flowing into the gas-liquid separator for: ;

[0031] In the formula: This represents the total alkali solution flow rate;

[0032] The heat balance equation for the gas-liquid separator is: ;

[0033] In the formula: This refers to the alkaline solution in the gas-liquid separator and the total heat capacity of the device, i.e. ; This indicates the temperature of the alkaline solution exiting the gas-liquid separator; and The heat dissipation from the separator is calculated as follows: ; ;

[0034] In the formula: This represents the equivalent heat dissipation area of ​​the electrolytic cell. Indicates the blackness of the separator;

[0035] The temperature process of the heat exchanger is modeled as follows: ; ; ;

[0036] In the formula: The total heat capacity of the heat exchanger structure and the alkaline solution within it, i.e. ; For heat exchange area; The heat transfer coefficient; Logarithmic temperature difference; and These are the inlet and outlet temperatures of the cooling water, respectively. This represents the total heat capacity of the cooling coil and the cooling water within it, i.e. The heat capacity of cooling water satisfy , This represents the volume of cooling water in the coil. and These represent the specific heat capacity and density of cooling water, respectively. This refers to the cooling water flow rate.

[0037] Preferably, the defined multi-dimensional performance metrics include:

[0038] Startup time : ;

[0039] In the formula: The initial energization time of the electrolytic cell. The temperature after the tank reaches the rated steady-state temperature for the first time. Time;

[0040] Maximum safe gradeability : ;

[0041] In the formula: To find the maximum safe gradeability, It is the slope of the input power change;

[0042] The constraints on the safe upper limit of HTO concentration and average temperature after the tank are as follows: ;

[0043] In the formula: and These are the oxygen impurity concentration in the hydrogen gas at specific times and the average temperature after the tank. and This is the upper limit for safe operation;

[0044] HTO Volatility Index : ;

[0045] In the formula: , The standard deviation and mean of the HTO concentration values;

[0046] Temperature consistency :

[0047] In the formula: They are respectively the 1st to the 1st in the system The outlet temperature of the platform tank;

[0048] Low load running time : ;

[0049] In the formula, It is the start of low load. The moment when the HTO concentration first reaches or exceeds 2.0 vol%.

[0050] Preferably, the multi-objective optimization function of the multi-objective optimization model is: ;

[0051] In the formula: Let the decision variable represent the design capacity of the separator. These are the weighting coefficients. To normalize the startup time, the system's rapid response capability from a cold state to rated operating conditions is quantified; considering the nonlinearity of the startup process, an integral form is used for calculation. ;

[0052] In the formula: This is the initial temperature of the electrolytic cell; The average outlet temperature of the tank;

[0053] To penalize the HTO volatility index and its deviation from the safety threshold, a penalty term is introduced for deviations exceeding the safety threshold: ;

[0054] in, This is the end time of the dynamic simulation;

[0055] For overall low-load performance index: ;

[0056] In the formula: This represents the maximum temperature non-uniformity at the electrolytic cell outlet under low load. These are the weighting coefficients;

[0057] Inequality constraints include:

[0058] Separator capacity constraints: ;

[0059] in, , Upper and lower limits for separator capacity design;

[0060] Dynamic process HTO constraints: ;

[0061] in, This is the tolerance value;

[0062] Thermal dynamic stability constraints: ;

[0063] in, The maximum allowable deviation between the average temperature and the reference value;

[0064] The equality constraints include the HTO accumulation dynamic model and thermal inertia model of the N-in-1 alkaline electrolysis hydrogen production system.

[0065] Preferably, the solution of the multi-objective optimization model includes:

[0066] A population initialization strategy based on trajectory sensitivity analysis is proposed, which introduces a directional sampling strategy based on trajectory sensitivity on the basis of standard random initialization: ; ;

[0067] in, Represents the initially generated first generation Individual; This represents a random number that is uniformly distributed in the interval [0,1]. It is a directional perturbation term based on trajectory sensitivity; It is the scaling factor; Trajectory sensitivity, representing the objective function. For separator capacity The rate of change; This is the separator capacity reference value used when calculating sensitivity; This represents a random number that follows a standard normal distribution. , These are the upper and lower limits for the separator capacity design;

[0068] For each individual in the population The accurate fitness value is obtained by solving the heat-mass coupling dynamic model: ;

[0069] in, Represents an individual The fitness vector contains three objective function values. This corresponds to normalized startup time, HTO fluctuation and penalty, and overall low-load performance index;

[0070] Each objective function is solved by numerical integration: ; ; ;

[0071] in, Weighting coefficients representing temperature non-uniformity; Indicates at time The non-uniformity of the outlet temperature of each electrolytic cell; It is an indicator function, when The value is 1 when the time is right, and 0 otherwise.

[0072] An improved fast non-dominated sorting algorithm is adopted, taking into account the degree of constraint violation: ;

[0073] in, Represents an individual Dominant Individual , Represents an individual The degree of constraint violation;

[0074] Crowding calculation: ;

[0075] in, Represents an individual The degree of congestion; and They respectively represent the objective function Above, with the individual The function values ​​of the next and previous adjacent individuals; and Represents the objective function in the current non-dominated layer. The maximum and minimum values; This represents the weighting coefficient used to constrain the degree of violation. It is a constant;

[0076] Adaptive genetic operator design:

[0077] Adaptive adjustment of crossover probability: ;

[0078] in, Indicates the adaptive crossover probability; and These represent the maximum and minimum crossover probabilities, respectively. Indicates the current generation number; Indicates the maximum number of generations; Standard deviation representing population crowding; This is a threshold used to determine whether a population is too concentrated.

[0079] Simulate binary crossover operation: ; ;

[0080] in, This represents the capacity value of the offspring 1 produced through the crossover operation; and This represents the capacity values ​​of the two selected parent individuals; Indicates the cross-distribution factor; This represents a random number that is uniformly distributed in the interval [0,1]. This represents the crossover index, whose value controls the degree of similarity between offspring and parents;

[0081] Polynomial mutation operation: ; ;

[0082] in, This represents the individual capacity value after mutation; Indicates variable-length asynchronous operation; This represents the variation distribution index, and its value controls the magnitude of the variation length.

[0083] Elite Retention and Environmental Selection: ; ; ; ;

[0084] in, The number of non-dominated layers; Indicates the first Temporary population after generational merger; Indicates the first The parent population of the generation; Indicates the first Offspring populations produced through genetic manipulation; Indicates population size; Represents a set Size; Indicates population size; Indicates the first The new paternal population; Indicates the first Non-dominated layer; This represents the index of the last non-dominated layer that was partially selected when filling a new population; This represents the relative change in the hypervolume index, used to measure the convergence of the algorithm. Indicates the first Generation population The excess volume value; The threshold for the algorithm;

[0085] The final output is the Pareto optimal solution set. : ;

[0086] in, It is the first non-dominated layer;

[0087] The improved NSGA-II algorithm is used to solve the optimization problem. The specific solution steps include:

[0088] Step 1: Set algorithm parameters, including population size. Maximum number of generations cross probability Different probabilities Decision space Randomly generate the initial parent population Evolutionary algebra ;

[0089] Step 2, for the current population Each individual Input the preset wind and solar power fluctuation data ; Run the HTO accumulated dynamic model and thermal inertia model to perform dynamic simulation; obtain multi-objective function values ​​through dynamic simulation calculation. Simultaneously verify whether each individual satisfies the HTO constraint and thermal dynamic stability constraint, and use the penalty function method to process individuals that violate the constraints in order to guide the search toward the feasible region;

[0090] Step 3, for the current population Perform non-dominated sorting operations to divide the population into multiple non-dominated levels based on Pareto dominance relationships. ,in The layer represents the current optimal Pareto front; to maintain the diversity of the solution set distribution, the crowding degree of each individual in the same non-dominated layer is calculated;

[0091] Step 4: A binary tournament selection mechanism is adopted to prioritize the selection of superior individuals with high non-dominant levels and high crowding. The selected individuals are subjected to a simulated binary crossover (SBX) operation to generate new offspring individuals. To further enhance population diversity, the offspring individuals are subjected to a polynomial mutation (PM) operation to introduce appropriate perturbation to explore new search spaces.

[0092] Step 5, transfer the parent population and offspring populations produced through genetic manipulation For temporary population ; for merging populations Perform non-dominated ranking and crowding comparison again, and select the top performers based on the elite retention strategy. The best individuals constitute the next generation of the parent population. ;

[0093] Step 6, update the evolutionary generation. The termination condition is set to reaching the maximum number of generations. The Pareto front either satisfies the convergence criterion; if the termination condition is met, the current non-dominated solution set is output as the Pareto optimal configuration of the separator capacity; otherwise, return to step 2 to continue iterative optimization.

[0094] A second aspect of the present invention provides a capacity configuration device for a separator in a many-to-one alkaline electrolysis hydrogen production system based on an improved multi-objective optimization search algorithm, comprising:

[0095] Establishment unit for establishing a thermal-mass coupling dynamic model of N-in-1 alkaline electrolysis hydrogen production system, including dynamic model of oxygen impurity HTO accumulation in hydrogen and thermal inertia model;

[0096] Define the unit to define multi-dimensional performance indicators, including startup time, HTO fluctuation index, temperature consistency, and low-load running time;

[0097] The establishment unit is used to construct a multi-objective optimization model with separator capacity as the decision variable and minimizing start-up time and HTO fluctuation index and maximizing low-load operation time as optimization objectives.

[0098] The solution unit is used to solve the multi-objective optimization model using the improved non-dominated sorting genetic algorithm NSGA-II, obtain the Pareto optimal solution set of the separator capacity, and obtain the separator capacity configuration result.

[0099] As can be seen from the above technical solution, the separator capacity configuration method for a multi-to-one alkaline electrolysis hydrogen production system based on an improved multi-objective optimization search algorithm provided in this invention first establishes a dynamic model of the thermal-mass coupling of the N-in-1 alkaline electrolysis hydrogen production system, including a dynamic model of HTO accumulation in hydrogen and a thermal inertia model; defines multi-dimensional performance indicators, including start-up time, HTO fluctuation index, temperature consistency, and low-load operating time; uses separator capacity as the decision variable, and minimizes start-up time and HTO fluctuation index, and maximizes low-load operating time as optimization objectives, constructs a multi-objective optimization model; uses the improved non-dominated sorting genetic algorithm NSGA-II to solve the multi-objective optimization model, obtains the Pareto optimal solution set for the separator capacity, and derives the separator capacity configuration result. This invention, through precise modeling of the thermal-mass coupling dynamic process inside the hydrogen production system, uses an optimization algorithm to solve for the separator capacity range that meets the multi-objective performance requirements, ensuring that hydrogen impurities meet standards, the system operates stably and with uniform temperature, while also considering the system's flexibility, safety, and economy. Attached Figure Description

[0100] Figure 1 This is a schematic diagram of the process for configuring the capacity of the separator in a multi-to-one alkaline electrolysis hydrogen production system based on an improved multi-objective optimization search algorithm, according to the present invention.

[0101] Figure 2 This is a schematic diagram of the overall structure of an N-in-1 alkaline electrolysis hydrogen production system.

[0102] Figure 3 This is a schematic diagram of HTO curves for different capacities.

[0103] Figure 4 This is a schematic diagram of the average downstream temperature curves for different separator capacities.

[0104] Figure 5 This is a schematic diagram of the temperature uniformity curve.

[0105] Figure 6 This is a schematic diagram of the HTO curves under high and low load conditions. Detailed Implementation

[0106] The technical solution and effects of the present invention will be further described in detail below with reference to the accompanying drawings.

[0107] This invention proposes a separator capacity design method for an alkaline electrolysis hydrogen production system based on an improved multi-objective optimization search algorithm. First, a thermal-mass coupling dynamic model of the N-in-1 alkaline electrolysis hydrogen production system is established, including a dynamic model of oxygen impurity (HTO) accumulation in hydrogen and a system thermal inertia model. Then, multi-dimensional performance indicators such as start-up time, HTO fluctuation index, temperature uniformity, and low-load operating time are defined to quantify system flexibility. Next, using separator capacity as the decision variable, and minimizing start-up time and HTO fluctuation index while maximizing low-load operating time are the optimization objectives, a multi-objective optimization problem is constructed. Finally, an improved non-dominated sorting genetic algorithm (NSGA-II) is introduced to solve the problem, obtaining the Pareto optimal solution set for the separator capacity, providing a quantitative basis for system design. The specific process is as follows: Figure 1 As shown:

[0108] The first aspect of this invention provides a method for configuring the capacity of a separator in a many-to-one alkaline electrolysis hydrogen production system based on an improved multi-objective optimization search algorithm, comprising:

[0109] A dynamic model of thermal-mass coupling was established for an N-in-1 alkaline electrolysis hydrogen production system, including a dynamic model of the accumulation of oxygen impurity HTO in hydrogen and a thermal inertia model.

[0110] Define multi-dimensional performance metrics, including startup time, HTO fluctuation index, temperature consistency, and low-load running time;

[0111] Using separator capacity as the decision variable, and minimizing start-up time and HTO fluctuation index, and maximizing low-load operating time as the optimization objectives, a multi-objective optimization model is constructed.

[0112] The improved non-dominated sorting genetic algorithm NSGA-II is used to solve the multi-objective optimization model, obtain the Pareto optimal solution set of the separator capacity, and obtain the separator capacity configuration result.

[0113] The establishment of a thermal-mass coupling dynamic model for an N-in-1 alkaline electrolysis hydrogen production system includes:

[0114] An industrial-grade N-in-1 alkaline electrolysis hydrogen production system consists of N parallel electrolyzers, a shared separator, a heat exchanger, a circulating pump, etc., with the overall structure as follows: Figure 2 As shown.

[0115] 1.1 HTO Accumulation Dynamic Model: (1) (2) (3)

[0116] In the formula: This indicates the current flowing through the electrolysis chamber; Indicates the number of electrolysis chambers; It is Faraday efficiency; It is Faraday's constant; , , These represent the molar flow rates of hydrogen impurities, hydrogen, and oxygen at the gas phase outlet of the oxygen-alkali separator, respectively.

[0117] The accumulation process can be summarized in three stages: hydrogen impurities first enter the cathode from the anode; then, they are transferred from the electrolyzer to the liquid phase of the separator along with the alkali solution circulation; finally, they enter the gas phase of the separator under the action of gravity, according to the definition of the accumulation process. :

[0118] Specifically, hydrogen impurities first enter the anode half-cell of each electrolyzer through three pathways: alkaline solution circulation, diffusion, and convection. (4)

[0119] This indicates the molar flow rate of hydrogen impurities carried by the alkali solution circulation during the stage from the anode to the cathode: (5)

[0120] This indicates the molar flow rate of hydrogen impurities transferred via diffusion from the electrolyzer to the liquid phase of the separator during the alkali solution circulation: (6)

[0121] This indicates the molar flow rate of hydrogen impurities transferred via convection during the gas phase stage of the separator under gravity. (7)

[0122] In the formula, i represents the electrolytic cell number; This indicates the solubility of hydrogen in alkaline solutions; Indicates the flow rate of the alkali solution; Indicates the density of the alkali solution; This represents the diffusion coefficient of hydrogen. This represents the difference in hydrogen concentration between the hydrogen side and the oxygen side; Indicates the thickness of the diaphragm; This indicates the permeability of hydrogen through the membrane; This represents the voltage difference across the half-cell. Indicates the viscosity of the alkaline solution;

[0123] The state-space model of HTO transfer and accumulation process in the N-in-1 system is as follows: (8)

[0124] In the formula: , and These represent the number of moles of hydrogen impurities in the anode half-cell, the liquid phase of the separator, and the gas phase of the separator, respectively. , , These represent the molar flow rates of hydrogen impurities transferred from the anode half-cell to the separator liquid phase, from the separator liquid phase to the gas phase, and exiting the gas-liquid separator, respectively, satisfying the following: (9)

[0125] In the formula: Indicates the gas-liquid separation time constant; This indicates the volume of the gas phase portion of the gas-liquid separator; It is the ideal gas constant; This refers to the separator outlet temperature.

[0126] 1.2 Thermal inertia model of the system:

[0127] The thermal balance equation for the electrolytic cell is: (10)

[0128] In the formula: This represents the total heat capacity of the electrolytic cell and the alkaline solution within it, i.e. The heat capacity of the alkali solution is calculated as follows: ,in This represents the volume of the alkaline solution in the tank. This is the specific heat capacity of the alkaline solution; and These represent the temperatures of the alkaline solution flowing into and out of the i-th electrolytic cell, respectively. Indicates the heat generated by electrolysis; This indicates that the electrolytic cell dissipates heat to the environment;

[0129] The heat generated by the electrolytic cell consists of the heat of reaction and the ohmic heat generated by the bypass current. (11)

[0130] In the formula: It is the thermal neutral voltage; Indicates the number of electrolysis chambers; This refers to the cell voltage; Indicates current; Indicates Faraday efficiency;

[0131] Heat dissipation from the electrolytic cell to the environment Heat dissipation by convection With radiative heat dissipation composition: (12)

[0132] In the formula: This represents the equivalent heat dissipation area of ​​the electrolytic cell. The ambient temperature; Boltzmann's constant; Indicates the blackness of the electrolytic cell; The coefficient of heat transfer by natural convection;

[0133] The natural convection heat transfer coefficient is calculated from the size of the electrolytic cell and the temperature difference between it and the environment: (13)

[0134] In the formula: The diameter of the electrolytic cell;

[0135] Temperature of alkaline solution flowing into the gas-liquid separator for: (14)

[0136] In the formula: This represents the total alkali solution flow rate;

[0137] The heat balance equation for the gas-liquid separator is: (15)

[0138] In the formula: This refers to the alkaline solution in the gas-liquid separator and the total heat capacity of the device, i.e. ; This indicates the temperature of the alkaline solution exiting the gas-liquid separator; and The heat dissipation from the separator is calculated as follows: (16) (17)

[0139] In the formula: This represents the equivalent heat dissipation area of ​​the electrolytic cell. Indicates the blackness of the separator;

[0140] The temperature process of the heat exchanger is modeled as follows: (18) (19) (20)

[0141] In the formula: The total heat capacity of the heat exchanger structure and the alkaline solution within it, i.e. ; For heat exchange area; The heat transfer coefficient; Logarithmic temperature difference; and These are the inlet and outlet temperatures of the cooling water, respectively. This represents the total heat capacity of the cooling coil and the cooling water within it, i.e. The heat capacity of cooling water satisfy , This represents the volume of cooling water in the coil. and These represent the specific heat capacity and density of cooling water, respectively. This refers to the cooling water flow rate.

[0142] Preferably, defining multi-dimensional performance metrics includes:

[0143] Startup time : (twenty one)

[0144] In the formula: The initial energization time of the electrolytic cell. The temperature after the tank reaches the rated steady-state temperature for the first time. Time;

[0145] Maximum safe gradeability : (twenty two)

[0146] In the formula: To find the maximum safe gradeability, It is the slope of the input power change;

[0147] The constraints on the safe upper limit of HTO concentration and average temperature after the tank are as follows: (twenty three)

[0148] In the formula: and These are the oxygen impurity concentration in the hydrogen gas at specific times and the average temperature after the tank. and For safe operation, the upper limits are set at 2.0 vol% and 85℃, respectively.

[0149] HTO Volatility Index : (twenty four)

[0150] In the formula: , The standard deviation and mean of the HTO concentration values;

[0151] Temperature consistency : (25)

[0152] In the formula: They are respectively the 1st to the 1st in the system The outlet temperature of the platform tank;

[0153] Low load running time :

[0154] In the formula, It is the start of low load. The moment when the HTO concentration first reaches or exceeds 2.0 vol%.

[0155] Constructing a multi-objective optimization problem based on NSGA-II:

[0156] 2.1 Separator capacity design model based on multi-objective optimization:

[0157] With separator capacity as the core decision variable and the goal of maximizing the flexibility and ensuring safe and stable operation of the alkaline electrolysis hydrogen production system, the multi-objective optimization function of the multi-objective optimization model is as follows:

[0158] (26)

[0159] In the formula: Let the decision variable represent the design capacity of the separator. These are the weighting coefficients. To normalize the startup time, the system's rapid response capability from a cold state to rated operating conditions is quantified; considering the nonlinearity of the startup process, an integral form is used for calculation. (27)

[0160] In the formula: This is the initial temperature of the electrolytic cell; The average outlet temperature of the tank;

[0161] The HTO volatility index and safety deviation penalty are used to measure the volatility of HTO. A penalty term is introduced for deviations exceeding the safety threshold to ensure the safety of the design. (28)

[0162] in, This is the end time of the dynamic simulation;

[0163] To comprehensively evaluate the system's stable operating time and temperature consistency (uniformity) under low load, the following low-load performance index is used: (29)

[0164] In the formula: This represents the maximum temperature non-uniformity at the electrolytic cell outlet under low load. These are the weighting coefficients;

[0165] Inequality constraints include:

[0166] Separator capacity constraints: (30)

[0167] in, , Upper and lower limits for separator capacity design;

[0168] Dynamic process HTO constraints: (31)

[0169] in, This is a tolerance value, allowing for small, acceptable deviations in HTO concentration near the safety limit;

[0170] Thermal dynamic stability constraints: (32)

[0171] in, The maximum allowable deviation between the average temperature and the reference value;

[0172] By combining the above objective function and constraints, a complete multi-objective optimization model for the separator capacity of the alkaline electrolysis hydrogen production system is constructed as follows: (26) (33)

[0173] A multi-objective optimization solution method based on the improved NSGA-II:

[0174] 3.1 Population initialization strategy based on trajectory sensitivity analysis:

[0175] A population initialization strategy based on trajectory sensitivity analysis is proposed, which introduces a directional sampling strategy based on trajectory sensitivity on the basis of standard random initialization: (34) (35)

[0176] in, Represents the initially generated first generation Individual; This represents a random number that is uniformly distributed in the interval [0,1]. It is a directional perturbation term based on trajectory sensitivity; It is the scaling factor; Trajectory sensitivity, representing the objective function. For separator capacity The rate of change; This is the separator capacity reference value used when calculating sensitivity; This represents a random number that follows a standard normal distribution. , These are the upper and lower limits for the separator capacity design;

[0177] For each individual in the population The accurate fitness value is obtained by solving the heat-mass coupling dynamic model: (36)

[0178] in, Represents an individual The fitness vector contains three objective function values. This corresponds to normalized startup time, HTO fluctuation and penalty, and overall low-load performance index; The transpose symbol for a vector or matrix, here representing a column vector;

[0179] Each objective function is solved by numerical integration: (37) (38) (39)

[0180] in, Weighting coefficients representing temperature non-uniformity; Indicates at time The non-uniformity of the outlet temperature of each electrolytic cell; , These represent the start and end times of low-load operation, respectively. It is an indicator function, when The value is 1 when the time is right, and 0 otherwise.

[0181] An improved fast non-dominated sorting algorithm is adopted, taking into account the degree of constraint violation: (40)

[0182] in, Represents an individual Dominant Individual , Represents an individual The degree of constraint violation;

[0183] Crowding calculation: (41)

[0184] in, Represents an individual The degree of congestion; and They respectively represent the objective function Above, with the individual The function values ​​of the next and previous adjacent individuals; and Represents the objective function in the current non-dominated layer. The maximum and minimum values; This represents the weighting coefficient used to constrain the degree of violation. It is a very small constant used to prevent the denominator from being zero;

[0185] 3.2 Adaptive Genetic Operator Design:

[0186] Adaptive adjustment of crossover probability: (42)

[0187] in, Indicates the adaptive crossover probability; and These represent the maximum and minimum crossover probabilities, respectively. Indicates the current generation number; Indicates the maximum number of generations; Standard deviation representing population crowding; This is a threshold used to determine whether a population is too concentrated.

[0188] Simulate binary crossover operation: (43) (41)

[0189] in, This represents the capacity value of the offspring 1 produced through the crossover operation; and This represents the capacity values ​​of the two selected parent individuals; Indicates the cross-distribution factor; This represents a random number that is uniformly distributed in the interval [0,1]. This represents the crossover index, whose value controls the degree of similarity between offspring and parents;

[0190] Polynomial mutation operation: (44) (45)

[0191] in, This represents the individual capacity value after mutation; Indicates variable-length asynchronous operation; This represents the variation distribution index, and its value controls the magnitude of the variation length.

[0192] Elite Retention and Environmental Selection: (46) (47) (48) (49)

[0193] in, The number of non-dominated layers; Indicates the first Temporary population after generational merger; Indicates the first The parent population of the generation; Indicates the first Offspring populations produced through genetic manipulation; Indicates population size; Represents a set Size; Indicates population size; Indicates the first The new paternal population; Indicates the first Non-dominated layer; This represents the index of the last non-dominated layer that was partially selected when filling a new population; This represents the relative change in the hypervolume index, used to measure the convergence of the algorithm. Indicates the first Generation population The hypervolume value is used to comprehensively measure the quality of the solution set (including convergence and diversity). The threshold for the algorithm;

[0194] The final output is the Pareto optimal solution set. : (50)

[0195] in, It is the first non-dominated layer;

[0196] The improved NSGA-II algorithm is used to solve the optimization problem. The specific solution steps include:

[0197] Step 1: Set algorithm parameters, including population size. Maximum number of generations cross probability Different probabilities Decision space Randomly generate the initial parent population Evolutionary algebra ;

[0198] Step 2, for the current population Each individual Input the preset wind and solar power fluctuation data ; Run the HTO accumulated dynamic model and thermal inertia model to perform dynamic simulation; obtain multi-objective function values ​​through dynamic simulation calculation. Simultaneously verify whether each individual satisfies the HTO constraint and thermal dynamic stability constraint, and use the penalty function method to process individuals that violate the constraints in order to guide the search toward the feasible region;

[0199] Step 3, for the current population Perform non-dominated sorting operations to divide the population into multiple non-dominated levels based on Pareto dominance relationships. ,in The layer represents the current optimal Pareto front; to maintain the diversity of the solution set distribution, the crowding degree of each individual in the same non-dominated layer is calculated;

[0200] Step 4: A binary tournament selection mechanism is adopted to prioritize the selection of superior individuals with high non-dominant levels and high crowding. The selected individuals are subjected to a simulated binary crossover (SBX) operation to generate new offspring individuals. To further enhance population diversity, the offspring individuals are subjected to a polynomial mutation (PM) operation to introduce appropriate perturbation to explore new search spaces.

[0201] Step 5, transfer the parent population and offspring populations produced through genetic manipulation For temporary population ; for merging populations Perform non-dominated ranking and crowding comparison again, and select the top performers based on the elite retention strategy. The best individuals constitute the next generation of the parent population. ;

[0202] Step 6, update the evolutionary generation. The termination condition is set to reaching the maximum number of generations. The Pareto front either satisfies the convergence criterion; if the termination condition is met, the current non-dominated solution set is output as the Pareto optimal configuration of the separator capacity; otherwise, return to step 2 to continue iterative optimization.

[0203] A second aspect of the present invention provides a separator capacity configuration device for a multi-to-one alkaline electrolysis hydrogen production system based on an improved multi-objective optimization search algorithm, used to implement the aforementioned method for configuring the separator capacity of a multi-to-one alkaline electrolysis hydrogen production system based on an improved multi-objective optimization search algorithm. The device includes:

[0204] Establishment unit for establishing a thermal-mass coupling dynamic model of N-in-1 alkaline electrolysis hydrogen production system, including dynamic model of oxygen impurity HTO accumulation in hydrogen and thermal inertia model;

[0205] Define the unit to define multi-dimensional performance indicators, including startup time, HTO fluctuation index, temperature consistency, and low-load running time;

[0206] The establishment unit is used to construct a multi-objective optimization model with separator capacity as the decision variable and minimizing start-up time and HTO fluctuation index and maximizing low-load operation time as optimization objectives.

[0207] The solution unit is used to solve the multi-objective optimization model using the improved non-dominated sorting genetic algorithm NSGA-II, obtain the Pareto optimal solution set of the separator capacity, and obtain the separator capacity configuration result.

[0208] The following embodiment is provided to explain the method of the present invention:

[0209] This example uses a 5 MW, 4-in-1 alkaline electrolysis hydrogen production system as the research object, and applies the method proposed in this invention to design the separator capacity. To verify the effectiveness of the method, a 24-hour dynamic simulation analysis was conducted based on actual wind and solar power fluctuation data, with the simulation step size set to 1 minute.

[0210] This example demonstrates the construction of a thermal-mass coupling dynamic model of an alkaline electrolysis hydrogen production system within the Mathematica environment, and the solution is obtained by integrating the NSGA-II multi-objective optimization algorithm. The main equipment parameters are set with reference to typical industrial-grade electrolysis hydrogen production systems.

[0211] Table 1 lists a comparison of system operating indicators under different separator size schemes, including key parameters such as minimum load operating time, HTO fluctuation index, and start-up time. Figure 3 , 4 The results show a comparison of the system's dynamic response under different separator capacity schemes; Figure 5 This shows the temperature uniformity performance under different capacities; Figure 6The curves showing the HTO concentration changes under high and low loads for each scheme are shown.

[0212] Table 1. Operating parameters of many-to-one systems with different separator sizes

[0213] The results show that the separator capacity design method for alkaline electrolysis hydrogen production systems based on the improved multi-objective optimization search algorithm proposed in this invention can systematically evaluate the impact of separator size on the dynamic characteristics of the system, effectively balance the multiple objective relationships between start-up response speed, HTO safety control and low-load operation capability, and provide reliable technical support for the optimized design of separators in alkaline electrolysis hydrogen production systems under wind and solar fluctuation scenarios.

[0214] In summary, this invention establishes a thermal-mass coupling dynamic model for an N-in-1 alkaline electrolysis hydrogen production system, constructs a multi-dimensional flexibility index system including start-up time, HTO fluctuation index, temperature uniformity, and low-load operating time, and realizes a quantitative correlation between separator capacity and system dynamic performance; proposes a multi-objective optimization model with separator capacity as the decision variable, and uses an improved NSGA-II algorithm for solution, achieving coordinated optimization design of the system among rapid response, safe operation, and low-load adaptability; and designs a population initialization strategy and adaptive genetic operator based on trajectory sensitivity analysis, significantly improving the algorithm's convergence efficiency and solution set distribution uniformity, overcoming the shortcomings of traditional trial-and-error methods such as long design cycles and difficulty in handling multi-objective conflicts.

[0215] The present invention has the following advantages:

[0216] Currently, there is a lack of research on the quantitative relationship between separator capacity and system flexibility and safety in the integrated application of alkaline electrolysis hydrogen production systems in renewable energy. The separator capacity design method based on multi-objective optimization proposed in this patent can systematically quantify the impact of separator size on system dynamic performance, effectively balance the conflict between rapid start-up, HTO safety control and low-load operation capability, and provide a theoretical basis for the optimized design of the system under fluctuating power scenarios.

[0217] The calculation is efficient and accurate. This patent uses an improved NSGA-II algorithm to solve multi-objective optimization problems. It combines trajectory sensitivity analysis initialization strategy and adaptive genetic operator to significantly improve computational efficiency while ensuring the quality of solution set. It effectively overcomes the shortcomings of traditional trial and error methods, such as long design cycle and difficulty in handling multi-objective trade-offs.

[0218] This patent proposes a systematic design method for separator capacity based on an improved multi-objective optimization algorithm. By constructing a thermal-mass coupling dynamic model and a multi-dimensional flexibility index system, and employing an improved NSGA-II algorithm integrating trajectory sensitivity initialization and adaptive genetic operators for solution, it achieves a leap from traditional empirical methods to model-driven, quantitative optimization in separator capacity design. This provides a scientific theoretical basis and efficient computational tools for the precise design and performance trade-offs of alkaline electrolysis hydrogen production systems under fluctuating wind and solar power conditions. By using separator capacity as a decision variable for multi-objective optimization, it can directly output the Pareto optimal capacity solution set that achieves the best trade-off between objectives such as rapid start-up, HTO safety control, and low-load operation capability. This effectively overcomes the limitations of traditional trial-and-error methods, which have long design cycles and rely on subjective experience. It enables designers to accurately select the optimal capacity scheme according to actual needs, significantly improving the dynamic response capability of the system and optimizing investment costs while ensuring the safe and stable operation of the system.

[0219] The above-disclosed embodiments are merely preferred embodiments of the present invention and should not be construed as limiting the scope of the invention. Those skilled in the art will understand that implementing all or part of the above-described embodiments and making equivalent changes in accordance with the claims of the present invention are still within the scope of the invention.

Claims

1. A method for configuring the capacity of a separator in a many-to-one alkaline electrolysis hydrogen production system based on an improved multi-objective optimization search algorithm, characterized in that, include: A dynamic model of thermal-mass coupling was established for an N-in-1 alkaline electrolysis hydrogen production system, including a dynamic model of the accumulation of oxygen impurity HTO in hydrogen and a thermal inertia model. Define multi-dimensional performance metrics, including startup time, HTO fluctuation index, temperature consistency, and low-load running time; Using separator capacity as the decision variable, and minimizing start-up time and HTO fluctuation index, and maximizing low-load operating time as the optimization objectives, a multi-objective optimization model is constructed. The improved non-dominated sorting genetic algorithm NSGA-II is used to solve the multi-objective optimization model, obtain the Pareto optimal solution set of the separator capacity, and obtain the separator capacity configuration result.

2. The separator capacity configuration method for a many-to-one alkaline electrolysis hydrogen production system based on an improved multi-objective optimization search algorithm as described in claim 1, characterized in that, The establishment of the thermal-mass coupling dynamic model of the N-in-1 alkaline electrolysis hydrogen production system includes: The industrial-grade N-in-1 alkaline electrolysis hydrogen production system includes N parallel electrolyzers, a shared gas-liquid separator, a heat exchanger, and a circulating pump; HTO Accumulation Dynamic Model: ； ； In the formula: This indicates the current flowing through the electrolysis chamber; Indicates the number of electrolysis chambers; It is Faraday efficiency; It is Faraday's constant; , , These represent the molar flow rates of hydrogen impurities, hydrogen, and oxygen at the gas phase outlet of the oxygen-alkali separator, respectively. According to the definition of accumulation process : ； ； ； ； In the formula, i represents the electrolytic cell number; This indicates the molar flow rate of hydrogen impurities carried by the alkaline solution circulation during the stage from the anode to the cathode. This indicates the molar flow rate of hydrogen impurities transferred by diffusion from the electrolyzer to the liquid phase stage of the separator along with the alkali solution circulation. This represents the molar flow rate of hydrogen impurities transferred through convection mechanisms as they enter the gas phase stage of the separator under the influence of gravity. This indicates the solubility of hydrogen in alkaline solutions; Indicates the flow rate of the alkali solution; Indicates the density of the alkali solution; This represents the diffusion coefficient of hydrogen. This represents the difference in hydrogen concentration between the hydrogen side and the oxygen side; Indicates the thickness of the diaphragm; This indicates the permeability of hydrogen through the membrane; This represents the voltage difference across the half-cell. Indicates the viscosity of the alkaline solution; The state-space model of HTO transfer and accumulation process in the N-in-1 system is as follows: ； In the formula: , and These represent the number of moles of hydrogen impurities in the anode half-cell, the liquid phase of the separator, and the gas phase of the separator, respectively. , , These represent the molar flow rates of hydrogen impurities transferred from the anode half-cell to the separator liquid phase, from the separator liquid phase to the gas phase, and exiting the gas-liquid separator, respectively, satisfying the following: ； In the formula: Indicates the gas-liquid separation time constant; This indicates the volume of the gas phase portion of the gas-liquid separator; It is the ideal gas constant; This refers to the separator outlet temperature. The thermal inertia model of the system: The thermal balance equation for the electrolytic cell is: ； In the formula: This represents the total heat capacity of the electrolytic cell and the alkaline solution within it, i.e. The heat capacity of the alkali solution is calculated as follows: ,in This represents the volume of the alkaline solution in the tank. This is the specific heat capacity of the alkaline solution; and These represent the temperatures of the alkaline solution flowing into and out of the i-th electrolytic cell, respectively. Indicates the heat generated by electrolysis; This indicates that the electrolytic cell dissipates heat to the environment; The heat generated by the electrolytic cell consists of the heat of reaction and the ohmic heat generated by the bypass current. ； In the formula: It is the thermal neutral voltage; Indicates the number of electrolysis chambers; This refers to the cell voltage; Indicates current; Indicates Faraday efficiency; Heat dissipation from the electrolytic cell to the environment Heat dissipation by convection With radiative heat dissipation composition: ； In the formula: This represents the equivalent heat dissipation area of ​​the electrolytic cell. The ambient temperature; Boltzmann's constant; Indicates the blackness of the electrolytic cell; The coefficient of heat transfer by natural convection; The natural convection heat transfer coefficient is calculated from the size of the electrolytic cell and the temperature difference between it and the environment: ； In the formula: The diameter of the electrolytic cell; Temperature of alkaline solution flowing into the gas-liquid separator for: ； In the formula: This represents the total alkali solution flow rate; The heat balance equation for the gas-liquid separator is: ； In the formula: This refers to the alkaline solution in the gas-liquid separator and the total heat capacity of the device, i.e. ; This indicates the temperature of the alkaline solution exiting the gas-liquid separator; and The heat dissipation from the separator is calculated as follows: ； ; In the formula: This represents the equivalent heat dissipation area of ​​the electrolytic cell. Indicates the blackness of the separator; The temperature process of the heat exchanger is modeled as follows: ； ； ； In the formula: The total heat capacity of the heat exchanger structure and the alkaline solution within it, i.e. ; For heat exchange area; The heat transfer coefficient; Logarithmic temperature difference; and These are the inlet and outlet temperatures of the cooling water, respectively. This represents the total heat capacity of the cooling coil and the cooling water within it, i.e. The heat capacity of cooling water satisfy , This represents the volume of cooling water in the coil. and These represent the specific heat capacity and density of cooling water, respectively. This refers to the cooling water flow rate.

3. The separator capacity configuration method for a many-to-one alkaline electrolysis hydrogen production system based on an improved multi-objective optimization search algorithm as described in claim 2, characterized in that... The defined multi-dimensional performance metrics include: Startup time : ; In the formula: The initial energization time of the electrolytic cell. The temperature after the tank reaches the rated steady-state temperature for the first time. Time; Maximum safe gradeability : ; In the formula: To find the maximum safe gradeability, It is the slope of the input power change; The constraints on the safe upper limit of HTO concentration and average temperature after the tank are as follows: ； In the formula: and These are the oxygen impurity concentration in the hydrogen gas at specific times and the average temperature after the tank. and This is the upper limit for safe operation; HTO Volatility Index : ; In the formula: , The standard deviation and mean of the HTO concentration values; Temperature consistency : ; In the formula: They are the 1st to the 1st in the system. The outlet temperature of the platform tank; Low load running time : ; In the formula, It is the start of low load. The moment when the HTO concentration first reaches or exceeds 2.0 vol%.

4. The separator capacity configuration method for a many-to-one alkaline electrolysis hydrogen production system based on an improved multi-objective optimization search algorithm as described in claim 3, characterized in that: The multi-objective optimization function of the multi-objective optimization model is: ； In the formula: Let the decision variable represent the design capacity of the separator. These are the weighting coefficients. To normalize the startup time, the system's rapid response capability from a cold state to rated operating conditions is quantified; considering the nonlinearity of the startup process, an integral form is used for calculation. ； In the formula: This is the initial temperature of the electrolytic cell; The average outlet temperature of the tank; To penalize the HTO volatility index and its deviation from the safety threshold, a penalty term is introduced for deviations exceeding the safety threshold: ； in, This is the end time of the dynamic simulation; For overall low-load performance index: ; In the formula: This represents the maximum temperature non-uniformity at the electrolytic cell outlet under low load. These are the weighting coefficients; Inequality constraints include: Separator capacity constraints: ; in, , Upper and lower limits for separator capacity design; Dynamic process HTO constraints: ; in, This is the tolerance value; Thermal dynamic stability constraints: ; in, The maximum allowable deviation between the average temperature and the reference value; The equality constraints include the HTO accumulation dynamic model and thermal inertia model of the N-in-1 alkaline electrolysis hydrogen production system.

5. The separator capacity configuration method for a many-to-one alkaline electrolysis hydrogen production system based on an improved multi-objective optimization search algorithm as described in claim 4, characterized in that... The solution to the multi-objective optimization model includes: A population initialization strategy based on trajectory sensitivity analysis is proposed, which introduces a directional sampling strategy based on trajectory sensitivity on the basis of standard random initialization: ； ； in, Represents the initially generated first generation Individual; This represents a random number that is uniformly distributed in the interval [0,1]. It is a directional perturbation term based on trajectory sensitivity; It is the scaling factor; Trajectory sensitivity, representing the objective function. For separator capacity The rate of change; This is the separator capacity reference value used when calculating sensitivity; This represents a random number that follows a standard normal distribution. , These are the upper and lower limits for the separator capacity design; For each individual in the population The accurate fitness value is obtained by solving the heat-mass coupling dynamic model: ； in, Represents an individual The fitness vector contains three objective function values. This corresponds to normalized startup time, HTO fluctuation and penalty, and overall low-load performance index; Each objective function is solved by numerical integration: ； ； ； in, Weighting coefficients representing temperature non-uniformity; Indicates at time The non-uniformity of the outlet temperature of each electrolytic cell; It is an indicator function, when The value is 1 when the time is right, and 0 otherwise. An improved fast non-dominated sorting algorithm is adopted, taking into account the degree of constraint violation: ； in, Represents an individual Dominant Individual , Represents an individual The degree of constraint violation; Crowding calculation: ； in, Represents an individual The degree of congestion; and They respectively represent the objective function Above, with the individual The function values ​​of the next and previous adjacent individuals; and Represents the objective function in the current non-dominated layer. The maximum and minimum values; This represents the weighting coefficient used to constrain the degree of violation. It is a constant; Adaptive genetic operator design: Adaptive adjustment of crossover probability: ; in, Indicates the adaptive crossover probability; and These represent the maximum and minimum crossover probabilities, respectively. Indicates the current generation number; Indicates the maximum number of generations; Standard deviation representing population crowding; This is a threshold used to determine whether a population is too concentrated. Simulate binary crossover operation: ； ； in, This represents the capacity value of the offspring 1 produced through the crossover operation; and This represents the capacity values ​​of the two selected parent individuals; Indicates the cross-distribution factor; This represents a random number that is uniformly distributed in the interval [0,1]. This represents the crossover index, whose value controls the degree of similarity between offspring and parents; Polynomial mutation operation: ； ； in, This represents the individual capacity value after mutation; Indicates variable-length asynchronous operation; This represents the variation distribution index, and its value controls the magnitude of the variation length. Elite Retention and Environmental Selection: ； ； ； ； in, The number of non-dominated layers; Indicates the first Temporary population after generational merger; Indicates the first The parent population of the generation; Indicates the first Offspring populations produced through genetic manipulation; Indicates population size; Represents a set Size; Indicates population size; Indicates the first The new paternal population; Indicates the first Non-dominated layer; This represents the index of the last non-dominated layer that was partially selected when filling a new population; This represents the relative change in the hypervolume index, used to measure the convergence of the algorithm. Indicates the first Generation population The excess volume value; The threshold for the algorithm; The final output is the Pareto optimal solution set. : ; in, It is the first non-dominated layer; The improved NSGA-II algorithm is used to solve the optimization problem. The specific solution steps include: Step 1: Set algorithm parameters, including population size. Maximum number of generations cross probability Different probabilities Decision space Randomly generate the initial parent population Evolutionary algebra ; Step 2, for the current population Each individual Input the preset wind and solar power fluctuation data ; Run the HTO accumulated dynamic model and thermal inertia model to perform dynamic simulation; obtain multi-objective function values ​​through dynamic simulation calculation. Simultaneously verify whether each individual satisfies the HTO constraint and thermal dynamic stability constraint, and use the penalty function method to process individuals that violate the constraints in order to guide the search toward the feasible region; Step 3, for the current population Perform non-dominated sorting operations to divide the population into multiple non-dominated levels based on Pareto dominance relationships. ,in The layer represents the current optimal Pareto front; to maintain the diversity of the solution set distribution, the crowding degree of each individual in the same non-dominated layer is calculated; Step 4: A binary tournament selection mechanism is adopted to prioritize the selection of superior individuals with high non-dominant levels and high crowding. The selected individuals are subjected to a simulated binary crossover (SBX) operation to generate new offspring individuals. To further enhance population diversity, the offspring individuals are subjected to a polynomial mutation (PM) operation to introduce appropriate perturbation to explore new search spaces. Step 5, transfer the parent population and offspring populations produced through genetic manipulation For temporary population ; for merging populations Perform non-dominated ranking and crowding comparison again, and select the top performers based on the elite retention strategy. The best individuals constitute the next generation of the parent population. ; Step 6, update the evolutionary generation. The termination condition is set to reaching the maximum number of generations. The Pareto front either satisfies the convergence criterion; if the termination condition is met, the current non-dominated solution set is output as the Pareto optimal configuration of the separator capacity; otherwise, return to step 2 to continue iterative optimization.

6. A separator capacity configuration device for a many-to-one alkaline electrolysis hydrogen production system based on an improved multi-objective optimization search algorithm, comprising: (The device performs the method according to any one of claims 1-5) Establishment unit for establishing a thermal-mass coupling dynamic model of N-in-1 alkaline electrolysis hydrogen production system, including dynamic model of oxygen impurity HTO accumulation in hydrogen and thermal inertia model; Define the unit to define multi-dimensional performance indicators, including startup time, HTO fluctuation index, temperature consistency, and low-load running time; The establishment unit is used to construct a multi-objective optimization model with separator capacity as the decision variable and minimizing start-up time and HTO fluctuation index and maximizing low-load operation time as optimization objectives. The solution unit is used to solve the multi-objective optimization model using the improved non-dominated sorting genetic algorithm NSGA-II, obtain the Pareto optimal solution set of the separator capacity, and obtain the separator capacity configuration result.

Images