Hydrogen fuel cell parameter optimization method and apparatus

Abstract

The application discloses a hydrogen fuel cell parameter optimization method and device, and relates to the technical field of fuel cells.The method comprises the following steps: establishing a multi-physical field coupling model according to the microstructure and macro-variable of a catalytic layer, which is used for quantifying the influence of platinum catalyst distribution on the performance of a fuel cell; determining an anisotropic platinum distribution function based on the concentration distribution characteristics of reactants and the multi-physical field coupling model; extracting multi-physical field indexes from the multi-physical field coupling model; constructing a proxy model between the platinum distribution function and the multi-physical field indexes based on a neural network; optimizing the platinum distribution parameters based on the proxy model through a multi-objective optimization algorithm; and screening the optimal anisotropic exponential platinum distribution parameter combination through a multi-attribute decision method. Through the multi-objective optimization algorithm and the multi-attribute decision method, the key multi-physical field indexes are collaboratively optimized, the optimal platinum distribution of the current fuel cell parameters can be accurately screened, and the solving precision performance of the optimal parameters of the hydrogen fuel cell is significantly improved.

Description

Technical Field

[0001] This application relates to the field of fuel cell technology, and in particular to methods and apparatus for optimizing parameters of hydrogen fuel cells. Background Technology

[0002] Proton exchange membrane fuel cells (PEMFCs) have become an important part of the future green energy architecture due to their high energy efficiency and cleanliness. However, due to the complex reactant transport and consumption mechanisms, the reactant concentration varies in different regions. Current industrial platinum deposition methods are still relatively crude, resulting in uniform or random platinum distribution in the finished catalyst. Therefore, efficient and reliable platinum distribution optimization strategies are urgently needed to improve platinum utilization.

[0003] Currently, research on platinum distribution in fuel cells mainly focuses on single distributions such as channel gradient, thickness gradient, and segmented step gradient, and the parameter definitions are highly dependent on expert experience, lacking interpretability. Furthermore, some "over-gas" optimization algorithms lack a global search of the parameter space. These problems result in insufficient accuracy in solving for the optimal parameters of current hydrogen fuel cells, leading to low overall fuel cell performance.

[0004] Therefore, improving the accuracy of solving the optimal parameters of hydrogen fuel cells is a problem that urgently needs to be solved. Summary of the Invention

[0005] The main objective of this application is to provide a method and apparatus for optimizing hydrogen fuel cell parameters, aiming to solve the technical problem of how to improve the accuracy of solving for the optimal parameters of hydrogen fuel cells.

[0006] To achieve the above objectives, this application proposes a method for optimizing hydrogen fuel cell parameters, the method comprising:

[0007] A multiphysics coupling model is established based on the microstructure and macroscopic variables of the catalyst layer. This multiphysics coupling model is used to quantify the influence of platinum catalyst distribution on battery performance.

[0008] Based on the reactant concentration distribution characteristics and the multiphysics coupling model, the anisotropic platinum distribution function is determined.

[0009] Multiphysics indices are extracted from the multiphysics coupling model, and a surrogate model between the platinum distribution function and the multiphysics indices is constructed based on a neural network.

[0010] Based on the aforementioned surrogate model, the platinum distribution parameters are optimized using a multi-objective optimization algorithm, and the optimal combination of anisotropic index platinum distribution parameters is selected using a multi-attribute decision method.

[0011] In one embodiment, the step of establishing a multiphysics coupling model based on the microstructure and macroscopic field variables of the catalyst layer includes:

[0012] An agglomerate model was established, and the thickness of the ionomer film and the thickness of the water film were calculated. The agglomerate model was used to represent the microstructure of the catalyst layer.

[0013] Based on the thickness of the ionomer film and the thickness of the water film, the local current density is calculated, which includes the cathode local current density and the anode local current density.

[0014] A multiphysics conservation equation set is established based on the macroscopic transport process of gas, liquid water, energy, and charge within the battery. The multiphysics conservation equation set includes mass, momentum, composition, water mode, and energy conservation equations.

[0015] The local current density at the cathode and the local current density at the anode are used as source terms and coupled into the multiphysics conservation equations. Through bidirectional data exchange between the macroscopic field variables and the aggregate model, a multiphysics coupled model is obtained.

[0016] In one embodiment, the step of determining the anisotropic platinum distribution function based on reactant concentration distribution characteristics and the multiphysics coupling model includes:

[0017] Based on the multiphysics coupling model, the mathematical relationship between reactant concentration distribution and reaction consumption rate is determined;

[0018] Based on the mathematical relationship between the reactant concentration distribution and the reaction consumption rate, a first function distribution form is determined to be adopted in the mass transfer direction of the flow channel and the catalyst layer thickness direction;

[0019] Based on the mathematical relationship between the reactant concentration distribution and the reaction consumption rate, a second function distribution form is determined to be adopted in the direction of the flow channel ribs;

[0020] Based on the first function distribution form and the second function distribution form, the anisotropic platinum distribution function is determined.

[0021] In one embodiment, the step of extracting multiphysics indices from the multiphysics coupling model and constructing a surrogate model between the platinum distribution function and the multiphysics indices based on a neural network includes:

[0022] Several performance indicators are obtained from the multiphysics coupling model, including gas distribution uniformity, liquid water saturation, output power, and temperature distribution uniformity.

[0023] The multiple performance indicators are weighted and fused to generate a multiphysics index;

[0024] Using the platinum distribution function as the input vector and the multiphysics index as the output vector, a multi-layer neural network is constructed.

[0025] The surrogate model is obtained by training the multilayer neural network based on the composite loss function.

[0026] In one embodiment, before the step of training the multilayer neural network based on a composite loss function to obtain the surrogate model, the following steps are included:

[0027] The monotonicity penalty loss function is obtained by penalizing the output difference that violates the preset monotonic direction through a linear rectification function.

[0028] The interval penalty loss function is obtained by penalizing the output difference that violates the preset interval through a linear rectification function;

[0029] A composite loss function is established based on the mean squared error loss function, the monotonicity penalty loss function, and the interval penalty loss function.

[0030] In one embodiment, the step of optimizing the platinum distribution parameters based on the surrogate model using a multi-objective optimization algorithm and selecting the optimal combination of anisotropic index platinum distribution parameters using a multi-attribute decision method includes:

[0031] Initialize the population, wherein each individual in the population is encoded to represent a set of platinum distribution parameters;

[0032] The multiphysics index value corresponding to each individual is calculated using the surrogate model;

[0033] Based on the non-dominated ordination results and crowding calculation results, individuals in the population are stratified and screened.

[0034] The offspring population is generated through selection, crossover, and mutation operations, and the population is iteratively updated by combining an elite retention strategy to obtain the Pareto solution set.

[0035] The optimal combination of anisotropic index platinum distribution parameters is selected from the Pareto solution set based on the multi-attribute decision method.

[0036] In one embodiment, the step of selecting the optimal combination of anisotropic index platinum distribution parameters from the Pareto solution set based on the multi-attribute decision method includes:

[0037] The multiphysics index values ​​corresponding to all solutions in the Pareto solution set are normalized to obtain the initial solution set;

[0038] Construct a weighted normalized decision matrix and determine the positive and negative ideal solutions for each indicator;

[0039] Calculate the Euclidean distance between each solution in the initial solution set and the positive ideal solution to obtain the first distance value;

[0040] Calculate the Euclidean distance between each solution in the initial solution set and the negative ideal solution to obtain the second distance value;

[0041] The relative proximity of each solution is calculated based on the first distance value and the second distance value, and the solution with the maximum proximity is selected as the optimal combination of anisotropic index platinum distribution parameters.

[0042] Furthermore, to achieve the above objectives, this application also proposes a hydrogen fuel cell parameter optimization device, which includes:

[0043] The coupling model building module is used to build a multi-physics coupling model based on the microstructure and macroscopic variables of the catalyst layer. The multi-physics coupling model is used to quantify the influence of platinum catalyst distribution on battery performance.

[0044] The platinum distribution determination module is used to determine the anisotropic platinum distribution function based on the reactant concentration distribution characteristics and the multiphysics coupling model.

[0045] The surrogate model construction module is used to extract multiphysics indices from the multiphysics coupling model and construct a surrogate model between the platinum distribution function and the multiphysics indices based on a neural network.

[0046] The optimal parameter selection module is used to optimize the platinum distribution parameters based on the surrogate model using a multi-objective optimization algorithm, and to select the optimal combination of anisotropic index platinum distribution parameters using a multi-attribute decision method.

[0047] In addition, to achieve the above objectives, this application also proposes a hydrogen fuel cell parameter optimization device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the hydrogen fuel cell parameter optimization method as described above.

[0048] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the hydrogen fuel cell parameter optimization method described above.

[0049] One or more technical solutions proposed in this application have at least the following technical effects:

[0050] (1) A multi-physics coupling model was established based on the microstructure and macroscopic variables of the catalyst layer. The multi-physics coupling model was used to quantify the influence of platinum catalyst distribution on battery performance. The multi-physics coupling model was adopted to establish a quantitative mapping relationship between platinum catalyst distribution and battery performance, overcoming the limitations of relying on expert experience in traditional methods.

[0051] (2) Based on the characteristics of reactant concentration distribution and multi-physics coupling model, the anisotropic platinum distribution function is determined, and the anisotropic platinum distribution parameters are generated through the characteristics of reactant concentration distribution, thus solving the problem of local reactant supply and demand imbalance caused by uniform distribution.

[0052] (3) Extract multi-physics field indices from the multi-physics field coupling model, and construct a proxy model between the platinum distribution function and the multi-physics field indices based on the neural network. The proxy model is constructed based on the neural network, and the optimization calculation is performed while retaining the correlation of the multi-physics field indices.

[0053] (4) Based on the surrogate model, the platinum distribution parameters are optimized using a multi-objective optimization algorithm, and the optimal combination of anisotropic index platinum distribution parameters is selected using a multi-attribute decision method. Through the multi-objective optimization algorithm and the multi-attribute decision method, the synergistic optimization of multiple physical field indices is achieved, which can efficiently and accurately select the optimal platinum distribution suitable for the current fuel cell parameters, and significantly improve the solution accuracy of the optimal parameters of the hydrogen fuel cell. Attached Figure Description

[0054] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0055] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 This is a flowchart illustrating the first embodiment of the hydrogen fuel cell parameter optimization method of this application;

[0057] Figure 2 This is a flowchart illustrating the second embodiment of the hydrogen fuel cell parameter optimization method of this application;

[0058] Figure 3 This is a schematic diagram of an aggregate model according to an embodiment of this application;

[0059] Figure 4 This is a flowchart illustrating the third embodiment of the hydrogen fuel cell parameter optimization method of this application;

[0060] Figure 5 This is a schematic diagram of the three-dimensional coordinate system of the catalyst layer in an embodiment of this application;

[0061] Figure 6 This is a flowchart illustrating the fourth embodiment of the hydrogen fuel cell parameter optimization method of this application;

[0062] Figure 7 This is a schematic diagram of the platinum distribution optimization framework in an embodiment of this application;

[0063] Figure 8 This is a schematic diagram of the module structure of the hydrogen fuel cell parameter optimization device according to an embodiment of this application;

[0064] Figure 9 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the hydrogen fuel cell parameter optimization method in this application embodiment.

[0065] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0066] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0067] To better understand the technical solution of this application, the following detailed description, in conjunction with the accompanying drawings, through specific embodiments and application scenarios, will illustrate the hydrogen fuel cell parameter optimization method, hydrogen fuel cell parameter optimization device, hydrogen fuel cell parameter optimization equipment, and storage medium provided in the embodiments of this application.

[0068] It should be noted that the hydrogen fuel cell parameter optimization method can be applied to the terminal, and can be executed by the hardware or software in the terminal.

[0069] The terminal includes, but is not limited to, portable communication devices such as mobile phones or tablets with touch-sensitive surfaces (e.g., touchscreen displays and / or touchpads). It should also be understood that, in some embodiments, the terminal may not be a portable communication device, but rather a desktop computer with touch-sensitive surfaces (e.g., touchscreen displays and / or touchpads).

[0070] The following embodiments describe a terminal including a display and a touch-sensitive surface. However, it should be understood that the terminal may include one or more other physical user interface devices such as a physical keyboard, mouse, and joystick.

[0071] Based on this, this application provides a method for optimizing hydrogen fuel cell parameters, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the hydrogen fuel cell parameter optimization method of this application.

[0072] In this embodiment, the hydrogen fuel cell parameter optimization method includes steps S10~S40:

[0073] Step S10: Establish a multiphysics coupling model based on the microstructure and macroscopic variables of the catalyst layer. The multiphysics coupling model is used to quantify the influence of platinum catalyst distribution on battery performance.

[0074] It should be noted that the microstructure of the catalyst layer can be a layered structure consisting of platinum particles, a carbon support, and an ionomer or water film encapsulating the catalyst. Macroscopic variables can be understood as physical quantities controlled by multiphysics conservation equations, such as gas velocity, pressure, liquid water saturation, and temperature, and can be coupled through a multiphysics coupling model that combines microscopic and macroscopic aspects.

[0075] Step S20: Based on the reactant concentration distribution characteristics and the multiphysics coupling model, determine the anisotropic platinum distribution function.

[0076] It should be noted that the reactant concentration distribution characteristics can represent the spatial distribution of reactants and are used to subsequently calculate the local reactant consumption rate. The anisotropic platinum distribution function is used to characterize the spatial non-uniformity of platinum loading.

[0077] Step S30: Extract multiphysics indices from the multiphysics coupling model and construct a surrogate model between the platinum distribution function and the multiphysics indices based on a neural network.

[0078] It should be noted that multiphysics indicators can be understood as performance evaluation parameters extracted from the coupled model, and the surrogate model can be understood as a multi-layer neural network model with platinum distribution parameters as input vectors and multiphysics indicators as output vectors.

[0079] Step S40: Based on the surrogate model, the platinum distribution parameters are optimized using a multi-objective optimization algorithm, and the optimal combination of anisotropic index platinum distribution parameters is selected using a multi-attribute decision method.

[0080] For example, a non-dominated sorting genetic algorithm can be used to perform multi-objective optimization of the platinum distribution, and a sorting method based on approximating the ideal solution can be used to select the parameter combination with the best comprehensive performance.

[0081] In this embodiment, a multiphysics coupling model was employed to establish a quantitative mapping relationship between platinum catalyst distribution and battery performance, overcoming the limitations of traditional methods that rely on expert experience. Anisotropic platinum distribution parameters were generated based on reactant concentration distribution characteristics, resolving the local reactant supply-demand imbalance problem caused by uniform distribution. A surrogate model was constructed based on a neural network, performing optimization calculations while preserving the correlation between multiphysics indices. Finally, through a multi-objective optimization algorithm and a multi-attribute decision-making method, collaborative optimization of multiphysics indices was achieved, efficiently and accurately selecting the optimal platinum distribution suitable for the current fuel cell parameters, significantly improving the accuracy of solving for the optimal parameters of the hydrogen fuel cell.

[0082] Reference Figure 2 , Figure 2 This is a flowchart illustrating the second embodiment of the hydrogen fuel cell parameter optimization method of this application, based on the above. Figure 1 The first embodiment shown illustrates a second embodiment of the hydrogen fuel cell parameter optimization method proposed in this application.

[0083] In the second embodiment, step S10 includes:

[0084] Step S101: Establish an agglomerate model and calculate the thickness of the ionomer film and the water film. The agglomerate model is used to represent the microstructure of the catalyst layer.

[0085] For example, such as Figure 3 This is a spherical agglomerate model of the catalyst layer in a proton exchange membrane fuel cell (PEMFC). The agglomerates consist of platinum particles and a carbon support. Because reactant gases cannot diffuse to the surface of the platinum particles inside the agglomerates, the agglomerates are divided into an outer effective region and an inner ineffective region. The agglomerate model can characterize the dissolution, diffusion, and adsorption processes of oxygen in a multiphase composite material composed of Pt. (Ionomer membrane thickness...) and water film thickness It can be calculated using formula (1).

[0086] (1)

[0087] In formula (1), It represents the volume fraction of the ionomer. This refers to the volume fraction of the platinum / carbon electrode. This represents the volume fraction of gas permeation through the diffusion layer. This represents the porosity of the catalyst layer. Furthermore... The volume fraction of the main pores occupied by the ionomer. It is the radius of the aggregate. It refers to the saturation level of liquid water.

[0088] Step S102: Calculate the local current density based on the ionomer film thickness and the water film thickness. The local current density includes the cathode local current density and the anode local current density.

[0089] For example, the local current density of the cathode is expressed as The local current density at the anode is expressed as It is expressed in the form of aggregate resistance terms and Butler-Volmer equation terms, as shown in formula (2).

[0090] (2)

[0091] In formula (2), It is specific surface area. It is the equivalent reaction coefficient. as well as It is the Henry's constant for oxygen and hydrogen. It is an aggregate equivalence factor. This is the reaction diffusion rate. The subscripts "a" and "c" represent parameters related to the anode and cathode, respectively.

[0092] Step S103: Establish a set of multiphysics conservation equations based on the macroscopic transport processes of gas, liquid water, energy, and charge within the battery.

[0093] It should be noted that the multiphysics conservation equations include mass, momentum, composition, water mode, and energy conservation equations.

[0094] Step S104: The local current density of the cathode and the local current density of the anode are used as source terms and coupled to the multiphysics conservation equations. Through bidirectional data exchange between the macroscopic field variables and the aggregate model, a multiphysics coupling model is obtained.

[0095] For example, based on the laws of conservation of mass, momentum, composition, modal water, and energy, the gas velocity in a fuel cell... ,pressure quality score Liquid water saturation fuel cell stack temperature It can be expressed by the following formula.

[0096]

[0097] in, , , , , and The porosity, viscosity, density, liquid water saturation, equivalent proton drag coefficient, and equivalent gas diffusion coefficient represent the reactant gas. , ,and , and These are the conservation source terms for mass, momentum, composition, modal water, and temperature. , , , , and These are the volume fraction and density of the ion membrane in the electrolyte, the equivalent mass of the dry membrane, and the specific heat capacity and equivalent heat capacity of the gas, liquid water, and gas. and These are the current density and the Faraday constant. and This represents the divergence and gradient calculations. This formula represents the multiphysics conservation equations, which, combined with formulas (1) and (2), constitute the multiphysics coupling model of the fuel cell. For example, high-fidelity data generated by COMSOL software is used, where the fuel cell multiphysics includes gas, water, electricity, and heat. The dataset parameter configuration is shown in Table 1.

[0098] Table 1

[0099]

[0100] In this embodiment, an agglomeration model was established and the thicknesses of the ionomer film and water film were calculated, quantifying the impact of platinum distribution on microscopic mass transfer resistance. Local current density was calculated based on film thickness, addressing the problem of traditional models neglecting the reaction gradient within the agglomerates. By establishing a set of conservation equations covering mass, momentum, composition, water, and energy, the macroscopic transport process was fully characterized. Finally, current density was coupled as a source term to the macroscopic equations, enabling bidirectional data exchange. This allows the multiphysics model to simultaneously reflect the interaction between the microscopic catalyst structure and macroscopic reactant transport, significantly improving prediction accuracy.

[0101] Reference Figure 4 , Figure 4 This is a flowchart illustrating the third embodiment of the hydrogen fuel cell parameter optimization method of this application, based on the above. Figure 2 The second embodiment shown presents a third embodiment of the hydrogen fuel cell parameter optimization method of this application.

[0102] In the third embodiment, step S20 includes:

[0103] Step S201: Determine the mathematical relationship between reactant concentration distribution and reaction consumption rate based on the multiphysics coupling model.

[0104] For example, the definition is as follows Figure 5The diagram shows a three-dimensional coordinate system, where the flow channel ribs, mass transfer, and catalyst thickness directions are defined as x, y, and z, respectively. Based on the aforementioned multiphysics coupling model, the reactant concentration can be determined. and consumption rate They exhibit a positive correlation, as shown in formula (3).

[0105] (3)

[0106] In formula (3) For reference current density, The concentration of reactants, For reference reactant concentrations. To activate loss The BV equation with as the independent variable.

[0107] Step S202: Based on the mathematical relationship between reactant concentration distribution and reaction consumption rate, determine the first function distribution form to be adopted in the mass transfer direction of the flow channel and the catalyst layer thickness direction.

[0108] It should be noted that the first function distribution can be a monotonically decreasing convex function distribution.

[0109] Step S203: Based on the mathematical relationship between reactant concentration distribution and reaction consumption rate, determine the second function distribution form to be adopted in the direction of the flow channel rib.

[0110] It should be noted that the second function distribution can be a symmetrical decreasing convex function distribution with the center line as the axis of symmetry.

[0111] Step S204: Determine the platinum distribution function for anisotropy based on the first function distribution form and the second function distribution form.

[0112] For example, higher reactant consumption rates are observed near the gas channel inlet and along the centerline, and proton accumulation near the cathode catalyst layer further exacerbates local consumption. To accommodate the spatial non-uniformity of transport paths and reaction consumption, a decreasing convex platinum distribution function in the y and z directions is proposed as a more suitable design strategy. In the z direction, the platinum distribution is designed as a decreasing convex function symmetrical about the centerline. Accordingly, the probability distribution function of the platinum catalyst is defined as shown in Equation (4).

[0113] (4)

[0114] In formula (4), , , These represent the attenuation parameters of the convex function in three directions, respectively. For platinum loading matching factor, and It is half the width of the catalyst layer.

[0115] In this embodiment, a mathematical relationship between concentration and consumption rate is established, revealing the key patterns of nonlinear decay of reactants along the flow channel and the low concentration trough in the central region of the ribs. A monotonically decreasing convex function is employed in the flow channel and thickness directions. A steep initial descent matches the demand in the high-reaction zone, while a gradual descent in the later stages avoids platinum waste in the low-concentration zone, ensuring that the loading distribution matches the slope of the consumption rate. A symmetrically decreasing convex function is employed in the rib direction. A central peak alleviates the mass transfer bottleneck, while gradual descent on both sides adapts to the diffusion advantages of the flow channel region. Finally, anisotropic parameters are generated through fusion, eliminating the linear errors of traditional uniform distribution.

[0116] Reference Figure 6 , Figure 6 This is a flowchart illustrating the fourth embodiment of the hydrogen fuel cell parameter optimization method of this application, based on the above. Figure 4 The third embodiment shown presents a third embodiment of the hydrogen fuel cell parameter optimization method of this application.

[0117] In the third embodiment, step S30 includes:

[0118] Step S301: Obtain multiple performance indicators from the multiphysics coupling model, including gas distribution uniformity, liquid water saturation, output power, and temperature distribution uniformity.

[0119] For example, four multiphysics indicators were extracted from the multiphysics coupling model: gas distribution uniformity (e.g., gas non-uniformity), liquid water saturation (e.g., liquid water accumulation), output power (e.g., maximum power), and temperature distribution uniformity (e.g., temperature non-uniformity). Among them, gas and temperature non-uniformity reflect the battery's durability, maximum power represents power generation efficiency, and liquid water accumulation is related to the stability and safety of operation.

[0120] Step S302: Weighted fusion of multiple performance indicators to generate multiphysics indicators.

[0121] For example, since Pt / C is prone to corrosion under localized high reaction concentrations, and hot spots are prone to irreversible battery damage, the non-uniformity index of gas concentration is defined by formula (5). Non-uniformity index of temperature .

[0122] (5)

[0123] In formula (5), It is the area of ​​the membrane electrode assembly. and These represent the local and average states (gas concentration and temperature) at the contact surfaces of the membrane and the cathode catalyst layer, respectively.

[0124] It should be noted that although various non-uniformities originate from different physical states, they are closely related to local reaction rates and exhibit similar trends. To facilitate subsequent surrogate modeling and multi-objective optimization, a weighted fusion non-uniformity index is introduced. The definition is as shown in formula (6).

[0125] (6)

[0126] In formula (6), This is a weighting coefficient for gas inhomogeneity. Due to concentration polarization limitations, the output power of a fuel cell does not increase indefinitely with increasing current density; its maximum power generally occurs in the concentration region. Therefore, the maximum power... Defined as an efficiency index, as in formula (7).

[0127] (7)

[0128] In formula (7), This refers to the fuel cell voltage, and its normal operating range is... ,and This represents the current as a function of voltage under the current platinum distribution. Water management in PEMFCs, especially the discharge of liquid water, is crucial for ensuring their operational stability and safety. Therefore, a water buildup index is defined. , is used to quantify the accumulation of liquid water saturation in the gas flow channel, gas diffusion layer and catalyst layer, and is expressed as in formula (8).

[0129] (8)

[0130] In formula (8), s represents the degree of saturation of liquid water. This represents the volume integral.

[0131] Step S303: Using the platinum distribution function as the input vector and the multiphysics index as the output vector, a multi-layer neural network is constructed.

[0132] It should be noted that multilayer neural networks are used to construct platinum distribution parameters and multiphysics indices. For example, the distribution parameters can be used... The input is a multi-physics performance index, and the output is a multi-physics performance index. Neural networks contain Hidden layers, whose parameters can be Each layer includes a weight matrix. and bias vector The model output is Formula (9) is obtained by calculating layer by layer through nonlinear transformation of each layer.

[0133] (9)

[0134] In formula (9), It is the first Layer output and the first Layer input. Activation function. This refers to the "Tanh" function. Neural networks can be implemented using Python, and the algorithm's hyperparameters are shown in Table 2.

[0135] Table 2

[0136]

[0137] Step S304: Train the multi-layer neural network model based on the composite loss function to obtain the surrogate model.

[0138] It should be noted that before step S304, the following steps are included: penalizing the output difference that violates the preset monotonic direction using a linear rectification function to obtain a monotonicity penalty loss function; penalizing the output difference that violates the preset interval using a linear rectification function to obtain an interval penalty loss function; and establishing a composite loss function based on the mean square error loss function, the monotonicity penalty loss function, and the interval penalty loss function.

[0139] For example, to guide a neural network to learn a predefined multiphysics monotonic relationship between input and output, the total loss function can be... Add monotonicity penalty item and interval penalty term , expressed as formula (10).

[0140] (10)

[0141] In formula (10), and It is a hyperparameter of monotonicity and interval penalty term.

[0142] For example, a set of input sample pairs can be constructed in the form of as well as ,in Determined by the platinum loading limit. Therefore, and The calculation method is as shown in formula (11).

[0143] (11)

[0144] In formula (11), It is a linear rectified function. Representing the The symbol weights of each output, and It is the first The upper and lower bounds of the output interval.

[0145] In this embodiment, indicators such as gas distribution uniformity, liquid water saturation, output power, and temperature distribution uniformity are extracted from a multiphysics coupling model, solving the problem of the one-sidedness of traditional single-indicator optimization. A unified multiphysics index is generated through weighted fusion, quantifying the synergistic impact of water management, output performance, and thermal safety. A neural network is constructed with platinum distribution parameters as input and fused indicators as output, replacing computationally intensive physical simulations. Finally, a composite loss function is used to train the surrogate model, achieving a balance between accuracy and efficiency.

[0146] In one implementation, based on the above embodiments and implementation methods, step S40 includes: initializing a population, where each individual in the population is encoded to represent a set of platinum distribution parameters; calculating the multiphysics index value corresponding to each individual using a surrogate model; stratifying and screening the individuals in the population based on the non-dominated ranking results and crowding calculation results; generating a progeny population through selection, crossover, and mutation operations, and iteratively updating the population using an elite retention strategy to obtain a Pareto solution set; and selecting the optimal combination of anisotropic index platinum distribution parameters from the Pareto solution set based on a multi-attribute decision method.

[0147] It should be noted that the step of selecting the optimal combination of anisotropic index platinum distribution parameters from the Pareto solution set based on the multi-attribute decision method includes: normalizing the multiphysics index values ​​corresponding to all solutions in the Pareto solution set to obtain an initial solution set; constructing a weighted normalized decision matrix and determining the positive and negative ideal solutions for each index; calculating the Euclidean distance between each solution in the initial solution set and the positive ideal solution to obtain a first distance value; calculating the Euclidean distance between each solution in the initial solution set and the negative ideal solution to obtain a second distance value; calculating the relative proximity of each solution based on the first and second distance values, and selecting the solution with the maximum proximity as the optimal combination of anisotropic index platinum distribution parameters.

[0148] For example, to optimize multi-physics performance indicators corresponding to different platinum distribution parameters, a non-dominated sorting genetic algorithm (NSGA-II) is used for multi-objective optimization based on a constructed multi-physics neural network surrogate model. This algorithm selects excellent solutions through Pareto sorting, maintains solution diversity using crowding distance, and introduces an elite retention mechanism to enhance convergence stability. Based on the optimal Pareto front, the technique for order preference by similarity to ideal solution (TOPSIS) is used to screen parameter combinations with optimal comprehensive performance, such as... Figure 7As shown, Figure 7 A schematic diagram of the platinum distribution optimization framework.

[0149] Specifically, it can generate a size of initial population Individual This represents a candidate solution and calculates its corresponding multi-objective output. Perform a non-dominated ranking; if the individual No inferior to in all objectives And superior to at least one objective Then it is called Dominate Calculate the crowding distance for each Pareto level, for individuals. Crowding degree is calculated by the difference between neighboring individuals on the objective function, and crowding distance is used to measure the sparsity of the solution. A binary tournament selection mechanism can be used to prioritize individuals with higher crowding degree and lower Pareto rank. Offspring individuals are generated by simulating binary crossover and polynomial operators, and the parent and offspring generations are merged into a population. For example, They are then non-dominated and ranked, and the top N individuals are selected based on rank and crowding level to form the next generation. The optimal Pareto front is generated after a specified number of iterations. Construct the decision matrix , Where i = 1 ~ m are candidate solutions, and j = 1 ~ n represent evaluation metrics. For the optimal Pareto front... Each indicator is dimensionless to eliminate dimensional differences between indicators. The normalized result is multiplied by the weights to obtain a weighted normalized matrix. The positive ideal solution represents the optimal value of each indicator, and the negative ideal solution represents the worst value of each indicator. Euclidean distance is used to calculate the relationship between each solution and both the positive and negative ideal solutions. The relative closeness of each solution is calculated based on the positive and negative distances, and the solutions are ranked accordingly. Finally, the solutions with the highest closeness are selected. Parameter combination This enables multi-objective collaborative optimization of multiple physics parameters.

[0150] In this embodiment, population coding is used to characterize platinum distribution parameters, transforming parameter optimization into a computable problem. A surrogate model is used to evaluate gas homogeneity, water saturation, and other indicators of each scheme within seconds, overcoming the hour-level bottleneck of traditional physical simulations. Pareto solutions are selected based on non-dominated ranking and crowding calculations, ensuring that the solutions simultaneously satisfy multi-objective non-dominance and diversity. Finally, a multi-attribute decision-making method is used to calculate the Euclidean distance between each solution and the positive and negative ideal solutions. The optimal platinum distribution parameters are automatically selected based on the criterion of maximum relative proximity, achieving synergistic optimization of improved gas homogeneity, reduced water saturation, and increased output power.

[0151] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the hydrogen fuel cell parameter optimization method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.

[0152] This application also provides a hydrogen fuel cell parameter optimization device; please refer to [reference needed]. Figure 8 The hydrogen fuel cell parameter optimization device includes:

[0153] The coupling model establishment module 10 is used to establish a multi-physics coupling model based on the microstructure and macroscopic variables of the catalyst layer. The multi-physics coupling model is used to quantify the influence of platinum catalyst distribution on battery performance.

[0154] The platinum distribution determination module 20 is used to determine the anisotropic platinum distribution function based on the reactant concentration distribution characteristics and the multiphysics coupling model.

[0155] The surrogate model construction module 30 is used to extract multiphysics indices from the multiphysics coupling model and construct a surrogate model between the platinum distribution function and the multiphysics indices based on a neural network.

[0156] The optimal parameter selection module 40 is used to optimize the platinum distribution parameters based on the surrogate model using a multi-objective optimization algorithm, and to select the optimal combination of anisotropic index platinum distribution parameters using a multi-attribute decision method.

[0157] The hydrogen fuel cell parameter optimization device provided in this application, employing the hydrogen fuel cell parameter optimization method described in the above embodiments, can solve the technical problem of how to improve the accuracy of solving for the optimal parameters of a hydrogen fuel cell. Compared with the prior art, the beneficial effects of the hydrogen fuel cell parameter optimization device provided in this application are the same as those of the hydrogen fuel cell parameter optimization method provided in the above embodiments, and other technical features in the hydrogen fuel cell parameter optimization device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0158] This application provides a hydrogen fuel cell parameter optimization device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the hydrogen fuel cell parameter optimization method in Embodiment 1 above.

[0159] The following is for reference. Figure 9The diagram illustrates a structural schematic suitable for implementing the hydrogen fuel cell parameter optimization device in the embodiments of this application. The hydrogen fuel cell parameter optimization device in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 9 The hydrogen fuel cell parameter optimization device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0160] like Figure 9 As shown, the hydrogen fuel cell parameter optimization device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the hydrogen fuel cell parameter optimization device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. The communication device 1009 allows the hydrogen fuel cell parameter optimization device to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 9 A hydrogen fuel cell parameter optimization device with various systems is shown; however, it should be understood that implementation or having all of the systems shown is not required. More or fewer systems may be implemented alternatively.

[0161] The hydrogen fuel cell parameter optimization device provided in this application, employing the hydrogen fuel cell parameter optimization method described in the above embodiments, can solve the technical problem of how to improve the accuracy of solving for the optimal parameters of a hydrogen fuel cell. Compared with the prior art, the beneficial effects of the hydrogen fuel cell parameter optimization device provided in this application are the same as those of the hydrogen fuel cell parameter optimization method provided in the above embodiments, and other technical features in this hydrogen fuel cell parameter optimization device are the same as those disclosed in the method of the previous embodiment, and will not be repeated here.

[0162] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0163] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the hydrogen fuel cell parameter optimization method in the above embodiments.

[0164] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, radio frequency (RF), etc., or any suitable combination thereof.

[0165] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0166] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described hydrogen fuel cell parameter optimization method, thereby solving the technical problem of how to improve the accuracy of solving for the optimal parameters of a hydrogen fuel cell. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the hydrogen fuel cell parameter optimization method provided in the above embodiments, and will not be repeated here.

[0167] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for optimizing parameters of a hydrogen fuel cell, characterized in that, The method includes: A multiphysics coupling model is established based on the microstructure and macroscopic variables of the catalyst layer. This multiphysics coupling model is used to quantify the influence of platinum catalyst distribution on battery performance. Based on the reactant concentration distribution characteristics and the multiphysics coupling model, the anisotropic platinum distribution function is determined. Multiphysics indices are extracted from the multiphysics coupling model, and a surrogate model between the platinum distribution function and the multiphysics indices is constructed based on a neural network. The multiphysics indices include gas concentration non-uniformity indices and temperature non-uniformity indices. Based on the aforementioned proxy model, the platinum distribution parameters are optimized using a multi-objective optimization algorithm, and the optimal combination of anisotropic index platinum distribution parameters is selected using a multi-attribute decision method. The step of establishing a multiphysics coupling model based on the microstructure and macroscopic field variables of the catalyst layer includes: An agglomerate model was established, and the thickness of the ionomer film and the thickness of the water film were calculated. The agglomerate model was used to represent the microstructure of the catalyst layer. Based on the thickness of the ionomer film and the thickness of the water film, the local current density is calculated, which includes the cathode local current density and the anode local current density. A multiphysics conservation equation set is established based on the macroscopic transport process of gas, liquid water, energy, and charge within the battery. The multiphysics conservation equation set includes mass, momentum, composition, water mode, and energy conservation equations. The local current density of the cathode and the local current density of the anode are used as source terms and coupled to the multiphysics conservation equations. Through bidirectional data exchange between the macroscopic field variables and the aggregate model, a multiphysics coupled model is obtained. The step of determining the anisotropic platinum distribution function based on the reactant concentration distribution characteristics and the multiphysics coupling model includes: Based on the multiphysics coupling model, the mathematical relationship between reactant concentration distribution and reaction consumption rate is determined; Based on the mathematical relationship between the reactant concentration distribution and the reaction consumption rate, a first function distribution form is determined to be adopted in the mass transfer direction of the flow channel and the catalyst layer thickness direction. The first function distribution form is a monotonically decreasing convex function distribution form. Based on the mathematical relationship between the reactant concentration distribution and the reaction consumption rate, a second function distribution form is determined to be adopted in the direction of the flow channel rib. The second function distribution form is a symmetrical decreasing convex function distribution form. Based on the first function distribution form and the second function distribution form, the anisotropic platinum distribution function is determined.

2. The method as described in claim 1, characterized in that, The step of extracting multiphysics indices from the multiphysics coupling model and constructing a surrogate model between the platinum distribution function and the multiphysics indices based on a neural network includes: Multiple performance indicators are obtained from the multiphysics coupling model, including the non-uniformity of gas concentration, liquid water saturation, output power, and temperature non-uniformity. The multiple performance indicators are weighted and fused to generate a multiphysics index; Using the platinum distribution function as the input vector and the multiphysics index as the output vector, a multi-layer neural network is constructed. The surrogate model is obtained by training the multilayer neural network based on the composite loss function.

3. The method as described in claim 2, characterized in that, Before the step of training the multi-layer neural network based on a composite loss function to obtain the surrogate model, the following steps are included: The monotonicity penalty loss function is obtained by penalizing the output difference that violates the preset monotonic direction through a linear rectification function. The interval penalty loss function is obtained by penalizing the output difference that violates the preset interval through a linear rectification function; A composite loss function is established based on the mean squared error loss function, the monotonicity penalty loss function, and the interval penalty loss function.

4. The method as described in claim 1, characterized in that, The steps of optimizing the platinum distribution parameters based on the surrogate model using a multi-objective optimization algorithm and selecting the optimal combination of anisotropic index platinum distribution parameters using a multi-attribute decision method include: Initialize the population, wherein each individual in the population is encoded to represent a set of platinum distribution parameters; The multiphysics index value corresponding to each individual is calculated using the surrogate model; Based on the non-dominated ordination results and crowding calculation results, individuals in the population are stratified and screened. The offspring population is generated through selection, crossover, and mutation operations, and the population is iteratively updated by combining an elite retention strategy to obtain the Pareto solution set. The optimal combination of anisotropic index platinum distribution parameters is selected from the Pareto solution set based on the multi-attribute decision method.

5. The method as described in claim 4, characterized in that, The step of selecting the optimal combination of anisotropic index platinum distribution parameters from the Pareto solution set based on the multi-attribute decision method includes: The multiphysics index values ​​corresponding to all solutions in the Pareto solution set are normalized to obtain the initial solution set; Construct a weighted normalized decision matrix and determine the positive and negative ideal solutions for each indicator; Calculate the Euclidean distance between each solution in the initial solution set and the positive ideal solution to obtain the first distance value; Calculate the Euclidean distance between each solution in the initial solution set and the negative ideal solution to obtain the second distance value; The relative proximity of each solution is calculated based on the first distance value and the second distance value, and the solution with the maximum proximity is selected as the optimal combination of anisotropic index platinum distribution parameters.

6. A hydrogen fuel cell parameter optimization device, characterized in that, The hydrogen fuel cell parameter optimization device includes: The coupling model building module is used to build a multi-physics coupling model based on the microstructure and macroscopic variables of the catalyst layer. The multi-physics coupling model is used to quantify the influence of platinum catalyst distribution on battery performance. The platinum distribution determination module is used to determine the anisotropic platinum distribution function based on the reactant concentration distribution characteristics and the multiphysics coupling model. The surrogate model construction module is used to extract multiphysics indices from the multiphysics coupling model and construct a surrogate model between the platinum distribution function and the multiphysics indices based on a neural network. The multiphysics indices include gas concentration non-uniformity indices and temperature non-uniformity indices. The optimal parameter selection module is used to optimize the platinum distribution parameters based on the surrogate model using a multi-objective optimization algorithm, and to select the optimal combination of anisotropic index platinum distribution parameters using a multi-attribute decision method. The coupling model establishment module is also used to establish an aggregate model and calculate the thickness of the ionomer film and the water film, wherein the aggregate model is used to represent the microstructure of the catalyst layer; based on the thickness of the ionomer film and the water film, the local current density is calculated, wherein the local current density includes the cathode local current density and the anode local current density; a multiphysics conservation equation set is established according to the macroscopic transport process of gas, liquid water, energy and charge in the battery, wherein the multiphysics conservation equation set includes mass, momentum, composition, water mode and energy conservation equations; the cathode local current density and the anode local current density are used as source terms and coupled to the multiphysics conservation equation set, and a multiphysics coupling model is obtained by bidirectional data exchange between the macroscopic field variables and the aggregate model; The platinum distribution determination module is further configured to: determine the mathematical relationship between reactant concentration distribution and reaction consumption rate based on the multiphysics coupling model; determine a first function distribution form in the mass transfer direction and catalyst layer thickness direction of the flow channel based on the mathematical relationship between reactant concentration distribution and reaction consumption rate, wherein the first function distribution form is a monotonically decreasing convex function distribution form; determine a second function distribution form in the flow channel rib direction based on the mathematical relationship between reactant concentration distribution and reaction consumption rate, wherein the second function distribution form is a symmetrically decreasing convex function distribution form; and determine an anisotropic platinum distribution function based on the first function distribution form and the second function distribution form.

7. A hydrogen fuel cell parameter optimization device, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the hydrogen fuel cell parameter optimization method as described in any one of claims 1 to 5.

8. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the hydrogen fuel cell parameter optimization method as described in any one of claims 1 to 5.

Images