Fuel cell membrane electrode performance evaluation method irrelevant to clamp runner

By setting multiple runner types and runner ratios, combined with neural network fitting technology, the problem of poor results in the existing test methods is solved, and the intuitive evaluation of membrane electrode performance is achieved, and the credibility of the test results is improved. and comparable.

CN119092750APending Publication Date: 2024-12-06NORTH CHINA ELECTRIC POWER UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411100476.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-12
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

Existing fuel cell membrane electrode performance testing methods rely on specific test fixtures and runner designs, resulting in poor comparability of results, and high-precision equipment and complex runner designs increase testing costs and are not suitable for large-scale performance evaluations.

Method used

By setting multiple runner types and runner ratios, a basic model set is constructed, and the current density and flow field uniformity are fitted using neural network and adam optimization algorithm to extract the slope set to achieve an intuitive reflection of the performance of membrane electrode materials. .

Benefits of technology

It improves the credibility and comparability of the test results, eliminates the impact of the fixture runner structure on the test results, reduces the testing cost, and makes the membrane electrode performance evaluation more repeatable and universal.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119092750A_ABST
    Figure CN119092750A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of fuel cells, and provides a fuel cell membrane electrode performance evaluation method irrelevant to a clamp runner, which comprises the following steps: constructing a fuel cell basic model set, dividing grids, setting simulation parameters and material attributes, simulating, acquiring a slope set and acquiring a target evaluation result. According to the invention, by setting a plurality of flow channel types and flow channel ratios, the influence of the flow channels on the performance of the fuel cell is eliminated, a test result which is more accurate and has higher repeatability on the membrane electrode is provided, the characteristics of the membrane electrode under different flow channel structures are predicted according to the obtained curve, and the accuracy of the test result is improved. A purer evaluation method is provided for development and optimization of an electrode material, and standardization of a fuel cell technology and deepening of basic research are promoted; the neural network is used for fitting, so that the slope is extracted, and the performance of the membrane electrode material is intuitively reflected.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of fuel cells, and in particular to a fuel cell membrane electrode performance evaluation method that is independent of a fixture flow channel. Background Art

[0002] As the core component of proton exchange membrane fuel cells, membrane electrodes determine the performance, efficiency and life of fuel cells, so better evaluation of membrane electrode performance is crucial in fuel cell research. However, current test methods rely on specific test fixtures and flow channel designs, which determine the distribution and flow characteristics of gases and liquids to a certain extent. Different research institutions and laboratories use different test fixtures and flow channel designs, resulting in poor comparability of test results. Even under the same operating conditions, there may be significant deviations in test results due to differences in fixture and flow channel design.

[0003] Different test systems may have differences in the control of temperature and humidity, which have a significant impact on the performance of fuel cells. Inconsistency in test conditions makes it difficult to compare results. The operating parameters used in the test process, such as pressure, gas flow rate, gas composition, etc., may vary depending on the equipment and experimental design, and these differences will affect the test results. At the same time, there are many different models and theories used to explain and simulate fuel cell performance. Different models may come to different conclusions, which affects the uniformity and consistency of the results. Due to the diversity of test equipment and methods, the results of the same experiment may vary greatly when repeated in different laboratories, affecting the credibility of the experimental results. Even within the same laboratory, there may be significant differences in the test results of different batches, resulting in poor reproducibility. At the same time, high-precision test equipment and complex flow channel design increase the cost of testing and are not suitable for large-scale performance evaluation. Summary of the invention

[0004] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a fuel cell membrane electrode performance evaluation method that is independent of the fixture flow channel, enrich the model types by setting a variety of flow channel types and flow channel ratios to improve the credibility of the results; and extract the slope by fitting using a neural network to intuitively reflect the performance of the membrane electrode material.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A fuel cell membrane electrode performance evaluation method independent of fixture flow channels, comprising:

[0007] A basic fuel cell model set is constructed based on four flow channel structures and multiple flow channel ratios;

[0008] Performing grid division on each partition of the fuel cell basic model set to obtain a fuel cell grid model set;

[0009] for setting simulation parameters and material properties for the fuel cell grid model set;

[0010] Simulating the fuel cell grid model set to obtain a fuel cell current density set and a membrane electrode flow field uniformity set;

[0011] Using a neural network and an adam optimization algorithm to fit the fuel cell current density set and the membrane electrode flow field uniformity set to obtain a slope set;

[0012] Different fuel cell membranes are evaluated according to the slope set to obtain target evaluation results.

[0013] Preferably, the four flow channel structures include: serpentine flow channels, straight parallel flow channels, interdigitated flow channels and grid-type flow channels.

[0014] Preferably, the fuel cell current density set and the membrane electrode flow field uniformity set are fitted using a neural network and an adam optimization algorithm to obtain a slope set, including:

[0015] Constructing a training set and a test set according to the fuel cell current density set and the membrane electrode flow field uniformity set;

[0016] Establishing a neural network model; the neural network model includes: an input layer, three linear hidden layers and an output layer; the input parameter of the input layer is the membrane electrode flow field uniformity data in the training set; the output parameter of the output layer is the fuel cell current density data in the training set;

[0017] The residual sum of squares is used as the loss function of the neural network model, and the neural network model is controlled to iterate to obtain the slope set.

[0018] Preferably, each partition of the fuel cell basic model set is grid-divided to obtain a fuel cell grid model set, including:

[0019] Importing the fuel cell basic model set into Hypermesh software for 2D and 3D mesh division to obtain an original mesh division model set;

[0020] Calculating a skew coefficient, an orthogonal quality coefficient, a mesh aspect ratio, and a mesh quality coefficient according to the original mesh partitioning model set to obtain a mesh partitioning quality index set;

[0021] Analyzing the grid partitioning quality index set according to a preset grid quality standard to obtain a partitioning evaluation result;

[0022] When the division evaluation result is that the division quality is unqualified, the original grid division model with unqualified division quality is re-gridded by changing the size, type and division method of the grid division to obtain the fuel cell grid model set.

[0023] Preferably, the material properties include: density, Cp specific heat, thermal conductivity, gas diffusivity and electrical conductivity.

[0024] Preferably, the three linear hidden layers include: a first linear hidden layer, a second linear hidden layer and a third linear hidden layer; the first linear hidden layer has 12 nodes; the second linear hidden layer has 4 nodes; and the third linear hidden layer has 1 node.

[0025] Preferably, the formula for obtaining the membrane electrode flow field uniformity set includes: as well as c v is the membrane electrode flow field uniformity data in the membrane electrode flow field uniformity set; σ is the standard deviation of the flow field on the contact surface of the fuel cell grid model; x is the average value of the flow field on the contact surface of the fuel cell grid model; n is the number of grid units on the contact surface of the fuel cell grid model; x i is the flow field value of the i-th grid unit on the contact surface of the fuel cell grid model.

[0026] Preferably, the partitions of the basic fuel cell model set include: bipolar plate, flow channel area and membrane electrode.

[0027] The present invention discloses the following technical effects:

[0028] The present invention provides a fuel cell membrane electrode performance evaluation method that is independent of fixture flow channels. By setting a variety of flow channel types and flow channel ratios, the problem of low result credibility is solved, and the model types are enriched. By using neural network fitting, the problem of slope extraction is solved, and an intuitive reflection of the performance of membrane electrode materials is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0030] Figure 1 A schematic diagram of a fuel cell membrane electrode performance evaluation process independent of a fixture flow channel provided by an embodiment of the present invention;

[0031] Figure 2 Schematic diagrams of four flow channel structures provided in embodiments of the present invention;

[0032] Figure 3 A curve diagram of the output current density and the flow field uniformity of the contact surfaces of each membrane electrode layer when the serpentine flow channel provided in an embodiment of the present invention adopts the third membrane electrode material;

[0033] Figure 4 A schematic diagram of the structure of a neural network model provided by an embodiment of the present invention;

[0034] Figure 5 The graphs of output current density and membrane electrode uniformity of the five materials provided in the embodiments of the present invention in different flow channels and different structural sizes are shown. DETAILED DESCRIPTION

[0035] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] The purpose of the present invention is to provide a fuel cell membrane electrode performance evaluation method that is independent of the fixture flow channel, enrich the model types by setting a variety of flow channel types and flow channel ratios to improve the credibility of the results; and extract the slope by fitting using a neural network to intuitively reflect the performance of the membrane electrode material.

[0037] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0038] Figure 1 A schematic diagram of a fuel cell membrane electrode performance evaluation process independent of a fixture flow channel provided by an embodiment of the present invention, such as Figure 1 As shown, the present invention provides a fuel cell membrane electrode performance evaluation method that is independent of the fixture flow channel, comprising:

[0039] Step 100: constructing a basic fuel cell model set according to four flow channel structures and multiple flow channel ratios;

[0040] Step 200: gridding each partition of the fuel cell basic model set to obtain a fuel cell grid model set;

[0041] Step 300: Setting simulation parameters and material properties for the fuel cell grid model set;

[0042] Step 400: simulating the fuel cell grid model set to obtain a fuel cell current density set and a membrane electrode flow field uniformity set;

[0043] Step 500: fitting the fuel cell current density set and the membrane electrode flow field uniformity set using a neural network and an adam optimization algorithm to obtain a slope set;

[0044] Step 600: Evaluate different fuel cell membranes according to the slope set to obtain target evaluation results.

[0045] Preferably, the four flow channel structures include: serpentine flow channels, straight parallel flow channels, interdigitated flow channels and grid-type flow channels.

[0046] Specifically, the fuel cell current density set and the membrane electrode flow field uniformity set are fitted using a neural network and an adam optimization algorithm to obtain a slope set, including:

[0047] Constructing a training set and a test set according to the fuel cell current density set and the membrane electrode flow field uniformity set;

[0048] Establishing a neural network model; the neural network model includes: an input layer, three linear hidden layers and an output layer; the input parameter of the input layer is the membrane electrode flow field uniformity data in the training set; the output parameter of the output layer is the fuel cell current density data in the training set;

[0049] The residual sum of squares is used as the loss function of the neural network model, and the neural network model is controlled to iterate to obtain the slope set.

[0050] Furthermore, each partition of the fuel cell basic model set is grid-divided to obtain a fuel cell grid model set, including:

[0051] Importing the fuel cell basic model set into Hypermesh software for 2D and 3D mesh division to obtain an original mesh division model set;

[0052] Calculating a skew coefficient, an orthogonal quality coefficient, a mesh aspect ratio, and a mesh quality coefficient according to the original mesh partitioning model set to obtain a mesh partitioning quality index set;

[0053] Analyzing the grid partitioning quality index set according to a preset grid quality standard to obtain a partitioning evaluation result;

[0054] When the division evaluation result is that the division quality is unqualified, the original grid division model with unqualified division quality is re-gridded by changing the size, type and division method of the grid division to obtain the fuel cell grid model set.

[0055] Specifically, the material properties include: density, Cp specific heat, thermal conductivity, gas diffusivity and electrical conductivity.

[0056] Optionally, the three linear hidden layers include: a first linear hidden layer, a second linear hidden layer and a third linear hidden layer; the first linear hidden layer has 12 nodes; the second linear hidden layer has 4 nodes; and the third linear hidden layer has 1 node.

[0057] Preferably, the formula for obtaining the membrane electrode flow field uniformity set includes: as well as c v is the membrane electrode flow field uniformity data in the membrane electrode flow field uniformity set; σ is the standard deviation of the flow field on the contact surface of the fuel cell grid model; x is the average value of the flow field on the contact surface of the fuel cell grid model; n is the number of grid units on the contact surface of the fuel cell grid model; x i is the flow field value of the i-th grid unit on the contact surface of the fuel cell grid model.

[0058] Specifically, the partitions of the basic fuel cell model set include: bipolar plate, flow channel area and membrane electrode.

[0059] Specifically, in order to increase the diversity of the flow channel structure of proton exchange membrane fuel cells, serpentine flow channels, straight parallel flow channels, interdigitated flow channels, and grid flow channels were selected as research objects, and Solidworks2020 was used for modeling. The model mainly includes bipolar plates (BP), flow channels, and membrane electrodes (membrane electrodes), and the membrane electrode includes a diffusion layer (GDL), a microporous layer (MPL), a catalytic layer (CL), and a proton exchange membrane (MEM). The size of BP is 50*50mm, the thickness of GDL is 0.2mm, the thickness of MPL is 0.02mm, the thickness of CL is 0.01mm, and the thickness of MEM is 0.05mm. Then, the ratio of the width of the flow channel to the ridge width is changed, and the depth of the flow channel is changed at the same time to increase the diversity of the flow channel structure. Each flow channel shape is transformed in the following way:

[0060] Channel width: ridge width = 1:1 (channel width 1 mm), changing the channel depth to 1.6 mm, 1.5 mm, 1.4 mm, 1.2 mm, 1 mm, 0.8 mm, 0.6 mm, 0.5 mm, 0.4 mm;

[0061] Channel width: ridge width = 1:2 (channel width 0.5 mm), change the channel depth, respectively 1.6 mm, 1.5 mm, 1.4 mm, 1.2 mm, 1 mm, 0.8 mm, 0.6 mm, 0.5 mm, 0.4 mm;

[0062] Channel width: ridge width = 2:1 (channel width 1 mm), changing the channel depth to 1.6 mm, 1.5 mm, 1.4 mm, 1.2 mm, 1 mm, 0.8 mm, 0.6 mm, 0.5 mm, 0.4 mm;

[0063] Channel width: ridge width = 1:2 (channel width 0.8mm), and the depth of the channel is changed to 1.6mm, 1.5mm, 1.4mm, 1.2mm, 1mm, 0.8mm, 0.6mm, 0.5mm, and 0.4mm respectively.

[0064] Figure 2 This is a schematic diagram of the structure with four types of flow channels, where the ratio of flow channel width to ridge width is 1:1, the flow channel width is 1mm, and the flow channel depth is 1mm. After the models of each part of each flow channel shape and each flow channel structure are built, they are combined and the combined model is saved in the format with the suffix .xt.

[0065] Furthermore, the constructed .xt file is imported into Hypermesh for 2D and 3D mesh division, and each part is meshed separately. The more appropriate mesh size is determined according to the size of each part. The quality of the mesh mainly depends on Skewness (skewness coefficient), Orthogonal Quality (orthogonal quality coefficient), Aspect Radio (mesh aspect ratio), and MeshMetric (mesh quality coefficient); the closer the Skewness value is to 0, the better the mesh quality is, the closer the OrthogonalQuality value is to 1, the better the mesh quality is, the closer the AspectRadio value is to 1, the better the mesh quality is, and the closer the MeshMetric value is to 1, the better the mesh quality is. If the obtained mesh quality is too different from the above-mentioned evaluation indicators, it needs to be re-divided. The size, type, and method of the mesh can be changed to obtain a model with better mesh quality. The output file is saved as a .cas file type.

[0066] Specifically, the meshed .cas file is imported into Fluent to set the operating parameters and material properties of the membrane electrode. The steps are as follows:

[0067] Enter define / mode / add in the console, press enter, select module 3, and open the proton exchange membrane fuel cell model;

[0068] Set the parameters for each part of penfc, check Multiphase, set Ref.Current Density to 7500, set Open-CircuitVoltage to 0.95, match the components of the anode in Anode, match the components of the cathode in Cathode, match the proton exchange membrane in Electrolyte, and set the parameters of each component. In the Report interface, set Electrolyte ProjectedArea to 0.0025m 2 , select terminal-a under Anode and terminal-c under Cathode; save the application and exit;

[0069] Set the unit area conditions, set plate-a and plate-c as solid areas, and the rest of the parts as fluid areas;

[0070] Boundary condition settings mainly include the inlet flow settings of inlet-a and inlet-c. Select mass-flow-inlet as the type, pressure-outlet as the type for outlet-a and outlet-c, and wall as the type for terminal-a and terminal-c. For specific boundary condition parameters, see Table 1.

[0071] Set the material properties of the membrane electrode, enter the material setting interface, and set the material properties of each part of the membrane electrode respectively. The specific parameters are shown in Table 2;

[0072] For the solver settings, select SIMPLE algorithm and adjust the sub-relaxation factor on the control page. Change the pressure to 0.7, the momentum to 0.3, and the rest to a smaller value, keeping it around 0.9.

[0073] On the calculation monitoring page, set the convergence standard and the maximum number of iterations;

[0074] For initialization settings, select standard initialization, select All-zones for calculation reference position, select Relative to unit area for reference system, and then set the number of iterations for calculation.

[0075] Table 1

[0076]

[0077] Table 2

[0078]

[0079]

[0080] Furthermore, the current density of the fuel cell under different flow channel structures and the uniformity of the flow field on the contact surface of each layer of the membrane electrode are simulated and calculated; the anode and cathode current densities of the control console are observed. When the error is less than 0.01% and the residual curve reaches the convergence standard, it is considered to have converged. The obtained current density is the stable output current density of the fuel cell, and the file is saved as .dat.h5 type;

[0081] Open the .dat.h5 file, find Results-Report-Surface Integral in the post-processing interface, select Uniformity Index-AreaWeighted as the report type, select Species-Molar Concentration ofh2 and Species-Molar Concentration ofo2 as the field variables when solving the uniformity of the concentration field, and select gdl-a / mpl-a, mpl-a / cl-a, cl-a / mem, gdl-c / mpl-c, mpl-c / cl-c, cl-c / mem as the Internal variables; when solving the uniformity of the velocity field, select Velocity as the field variable, and select gdl-a / mpl-a, mpl-a / cl-a, cl-a / mem, gdl-c / mpl-c, mpl-c / cl-c, cl-c / mem as the Internal variables; when solving the uniformity of the pressure field, select Velocity as the field variable, and select gdl-a / mpl-a, mpl-a / cl-a, cl-a / mem, gdl-c / mpl-c, mpl-c / cl-c, cl-c / mem as the Internal variables; Select Pressure as the variable, and select gdl-a / mpl-a, mpl-a / cl-a, cl-a / mem, gdl-c / mpl-c, mpl-c / cl-c, cl-c / mem as Internal respectively; when solving the uniformity of the temperature field, select Temperature as the field variable, and select gdl-a / mpl-a, mpl-a / cl-a, cl-a / mem, gdl-c / mpl-c, mpl-c / cl-c, cl-c / mem as Internal respectively. Figure 3 The ratio of the channel width to the ridge width of the serpentine channel is 1:1, the channel width is 1mm, and the channel depths are 0.4mm, 0.5mm, 0.6mm, 0.8mm, 1mm, 1.2mm, 1.4mm, 1.5mm, and 1.6mm respectively. The relationship between the output current density and the flow field uniformity of the contact surfaces of each layer of the membrane electrode using the third membrane electrode material. Figure 3In the first figure, the horizontal coordinate Uniformity of membrane electrode contact surface concentration is the concentration field uniformity of the contact surface of each layer of MEA (membrane electrode), the vertical coordinate Current density is the output current density, gdl-a / mpl-a is the contact surface between the anode diffusion layer and the anode microporous layer, mpl-a / cl-a is the contact surface between the anode microporous layer and the anode catalyst layer, and cl-a / mem is the contact surface between the anode catalyst layer and the proton exchange membrane; Figure 3 In the second figure, the horizontal coordinate Velocity uniformity of each layer of membrane electrode is the velocity field uniformity of the contact surface of each layer of MEA, the vertical coordinate Current density is the output current density, gdl-a / mpl-a is the contact surface between the anode diffusion layer and the anode microporous layer, mpl-a / cl-a is the contact surface between the anode microporous layer and the anode catalyst layer, and cl-a / mem is the contact surface between the anode catalyst layer and the proton exchange membrane; Figure 3 In the third figure, the horizontal coordinate Pressure uniformity of each layer of membrane electrode is the pressure field uniformity of the contact surface of each layer of MEA, the vertical coordinate Current density is the output current density, gdl-a / mpl-a is the contact surface between the anode diffusion layer and the anode microporous layer, mpl-a / cl-a is the contact surface between the anode microporous layer and the anode catalyst layer, and cl-a / mem is the contact surface between the anode catalyst layer and the proton exchange membrane; Figure 3 In the fourth figure, the horizontal axis Temperature uniformity of each layer of membrane electrode is the temperature field uniformity of the contact surfaces of each layer of MEA, the vertical axis Current density is the output current density, gdl-a / mpl-a is the contact surface between the anode diffusion layer and the anode microporous layer, mpl-a / cl-a is the contact surface between the anode microporous layer and the anode catalyst layer, and cl-a / mem is the contact surface between the anode catalyst layer and the proton exchange membrane.

[0082] Preferably, the fuel cell current density and the membrane electrode flow field uniformity are linearly fitted; considering that there are many parameters for the flow field uniformity of the contact surfaces of each layer of the membrane electrode, in order to find the linear relationship between the output current density and the uniformity of the membrane electrode of the proton exchange membrane fuel cell under different flow channels and different structural sizes, it is transformed into a neural network model with learnable weights, and the linear relationship between the current density and the membrane electrode uniformity is fitted through the Adam optimization algorithm. The final slope is the evaluation standard for different membrane electrode materials.

[0083] Specifically, the neural network model uses three linear hidden layers, and the loss function uses the residual sum of squares. Each input layer is the flow field uniformity of the membrane electrode contact surface, with a total of 24 parameters, and the output layer is the current density. From the input layer to the output layer, there are three linear hidden layers in the middle. The first hidden layer has 12 nodes, the second hidden layer has 4 nodes, and the third hidden layer has 1 node. Each node has a skewness and a weight for each input value to the node, that is, a value f(z) can be obtained at each node, and this value is used as the input value of the next layer of neurons until the output is established and a relationship corresponding to the current density is established. Figure 4 , where x 1 , x 2 , x 3 ...is the uniformity of each flow field on the contact surface of each layer of MEA, that is, the input value (input), w 1 , w 2 , w 3 …is the weight of each input value to the node, that is, Weights. In order to map the input of the neuron to the function of the output end in the neural network, the activation function is introduced. Output is the output value. The value obtained at each node is Where b is the skewness under each node, w i is the weight of the ith node.

[0084] refer to Figure 5 , under the same weight and skewness, the output current density and membrane electrode uniformity of the other four membrane electrode materials in different flow channels and different structural sizes also show a linear relationship. The output current density and membrane electrode uniformity of the proton exchange membrane fuel cell with five membrane electrode materials in different flow channels and different structural sizes all have linear expressions. The horizontal coordinate Uniformity of membrane electrode is the uniformity of MEA, and the vertical coordinate Current density is the current density: y1 = 0.693x-0.320, y2 = 0.671x-0.272, y3 = 1.210x-0.485, y4 = 0.791x-0.102, y5 = 0.681x-0.186; the squares of the correlation coefficients (r2) are 0.83, 0.83, 0.82, 0.91, and 0.89, respectively. At the same time, it is found that the slopes of the linear relationship between the output current density and the membrane electrode obtained by using the first material and the fifth material are close to the same, so in Figure 5The two lines overlap, and its slope K reflects the material properties of the membrane electrode itself. It is not difficult to find from Table 2 that the various properties of the various parts of the materials of the first membrane electrode and the various properties of the various parts of the materials of the fifth membrane electrode are very small, so the slope K obtained by linear fitting is almost the same. When the membrane electrode adopts the third material, its slope K is larger. When comparing the various properties of the third material in Table 2 with the various properties of the other four materials, it is found that the conductivity of the proton exchange membrane (MEM) in the third membrane electrode is much higher than the conductivity of MEM in the other four. Considering that high proton conductivity can ensure the rapid transfer of protons in the membrane and improve the efficiency of the battery, the slope K is larger. When the membrane electrode adopts the fourth material, the slope K obtained is second only to the K obtained when the membrane electrode adopts the third material. From the comparison in Table 2, it is found that the various properties of the catalyst layer (CL) in the fourth membrane electrode material are higher than the various properties of CL in the other four membrane electrode materials, mainly manifested in high thermal conductivity and electrical conductivity, which also significantly improves the oxidation and oxygen reduction reaction rates of the fuel, thereby improving the output current density of the poor fuel cell.

[0085] Furthermore, for each membrane electrode material, the output current density under different flow channel structure sizes and the uniformity of the corresponding membrane electrode show a linear relationship, and its slope K reflects the material properties of the membrane electrode itself. Based on the obtained linear expression, the output current density under other structure sizes can also be predicted. At the same time, by comparing the slope K under different membrane electrode materials, the performance of the membrane electrode material can be intuitively reflected.

[0086] The beneficial effects of the present invention are as follows:

[0087] The present invention provides a fuel cell membrane electrode performance evaluation method that is independent of fixture flow channels, takes serpentine flow channels, straight parallel flow channels, interdigital flow channels, and grid-type flow channels as research objects, increases the diversity of flow channel structures by changing the ratio of flow channel width to ridge width and the depth of the flow channels, and sets 5 kinds of membrane electrode material properties at the same time. From the perspective of uniformity, the uniformity of the flow field of concentration field, velocity field, pressure field, and temperature field at the contact surface of each layer of the membrane electrode is comprehensively considered, and the adam optimization algorithm in the neural network is used to fit the output current density of different flow channel shapes and different structural sizes under each membrane electrode material property. The linear relationship between the uniformity of the corresponding membrane electrode is obtained, and its slope K reflects the material properties of the membrane electrode itself. According to the linear expression obtained, the output current density under other structural sizes can be predicted, and the performance of the membrane electrode material can be intuitively reflected by comparing the slope K under different membrane electrode materials. The method eliminates the influence of the fixture flow channel structure on the membrane electrode test results, so that the test results mainly reflect the performance of the membrane electrode itself, rather than being affected by external test equipment, and greatly improves the repeatability and comparability of the test results.

[0088] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0089] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A fuel cell membrane electrode performance evaluation method independent of fixture flow channels, characterized in that: include: A basic fuel cell model set is constructed based on four flow channel structures and multiple flow channel ratios; Performing grid division on each partition of the fuel cell basic model set to obtain a fuel cell grid model set; for setting simulation parameters and material properties for the fuel cell grid model set; Simulating the fuel cell grid model set to obtain a fuel cell current density set and a membrane electrode flow field uniformity set; Using a neural network and an adam optimization algorithm to fit the fuel cell current density set and the membrane electrode flow field uniformity set to obtain a slope set; Different fuel cell membranes are evaluated according to the slope set to obtain target evaluation results.

2. A fuel cell membrane electrode performance evaluation method independent of fixture flow channels according to claim 1, characterized in that: The four flow channel structures include: serpentine flow channel, straight parallel flow channel, interdigitated flow channel and grid flow channel.

3. A fuel cell membrane electrode performance evaluation method independent of fixture flow channels according to claim 1, characterized in that: The fuel cell current density set and the membrane electrode flow field uniformity set are fitted using a neural network and an adam optimization algorithm to obtain a slope set, including: Constructing a training set and a test set according to the fuel cell current density set and the membrane electrode flow field uniformity set; Establishing a neural network model; the neural network model includes: an input layer, three linear hidden layers and an output layer; the input parameter of the input layer is the membrane electrode flow field uniformity data in the training set; the output parameter of the output layer is the fuel cell current density data in the training set; The residual sum of squares is used as the loss function of the neural network model, and the neural network model is controlled to iterate to obtain the slope set.

4. A fuel cell membrane electrode performance evaluation method independent of fixture flow channels according to claim 1, characterized in that: Each partition of the fuel cell basic model set is grid-divided to obtain a fuel cell grid model set, including: Importing the fuel cell basic model set into Hypermesh software for 2D and 3D mesh division to obtain an original mesh division model set; Calculating a skew coefficient, an orthogonal quality coefficient, a mesh aspect ratio, and a mesh quality coefficient according to the original mesh partitioning model set to obtain a mesh partitioning quality index set; Analyzing the grid partitioning quality index set according to a preset grid quality standard to obtain a partitioning evaluation result; When the division evaluation result is that the division quality is unqualified, the original grid division model with unqualified division quality is re-gridded by changing the size, type and division method of the grid division to obtain the fuel cell grid model set.

5. A fuel cell membrane electrode performance evaluation method independent of fixture flow channels according to claim 1, characterized in that: The material properties include density, Cp specific heat, thermal conductivity, gas diffusivity, and electrical conductivity.

6. A fuel cell membrane electrode performance evaluation method independent of fixture flow channels according to claim 1, characterized in that: The three linear hidden layers include: a first linear hidden layer, a second linear hidden layer and a third linear hidden layer; the first linear hidden layer has 12 nodes; the second linear hidden layer has 4 nodes; and the third linear hidden layer has 1 node.

7. A fuel cell membrane electrode performance evaluation method independent of fixture flow channels according to claim 1, characterized in that: The formula for obtaining the membrane electrode flow field uniformity set includes: as well as c v is the membrane electrode flow field uniformity data in the membrane electrode flow field uniformity set; σ is the standard deviation of the flow field on the contact surface of the fuel cell grid model; x is the average value of the flow field on the contact surface of the fuel cell grid model; n is the number of grid units on the contact surface of the fuel cell grid model; x i is the flow field value of the i-th grid unit on the contact surface of the fuel cell grid model.

8. A fuel cell membrane electrode performance evaluation method independent of fixture flow channels according to claim 1, characterized in that: The partitions of the basic fuel cell model set include: bipolar plate, flow channel area and membrane electrode.