Fuel cell physical field simulation method, system and device

Abstract

The invention provides a fuel cell physical field simulation method, system and device, and relates to the technical field of fuel cell physical field simulation. Constructing a PINN model: taking the three-dimensional coordinate value and time of the galvanic pile temperature field as input parameters; taking the temperature in the electric pile as an output result; defining a stack heat conduction formula and constructing a residual function; taking the initial temperature distribution and the boundary condition as constraint conditions; defining a loss function based on the residual function and the constraint condition; training and evaluating the PINN model; and reconstructing a temperature field through the PINN model. Real-time sensor data can be combined, physical constraints and limited data points are combined, and accurate physical field distribution can still be obtained under the condition of insufficient data; according to the scheme, accurate and dynamic physical field prediction can be provided, the calculation complexity can be reduced, the efficiency is higher when the physical field is reconstructed, and the calculation amount is remarkably reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of fuel cell physical field simulation, and in particular to a fuel cell physical field simulation method, system and device. Background Art

[0002] As an efficient and clean energy conversion device, proton exchange membrane fuel cell (PEMFC) has broad application prospects in the fields of automobiles, distributed energy systems, etc. However, the performance and life of PEMFC are affected by many factors, among which physical fields such as temperature distribution play a key role in the operation stability and degradation mechanism of fuel cells. In particular, local overheating or uneven temperature of fuel cells can lead to deactivation of the catalyst layer, membrane drying or flooding problems, which in turn affect the overall performance of the battery. Therefore, accurately predicting and simulating physical fields such as temperature distribution inside PEMFC is crucial to optimize its thermal management design and extend its service life.

[0003] At present, relevant solutions have emerged to address the above issues, but there are still some shortcomings: The patent application with the publication number CN110336057A, entitled A method for constructing a two-dimensional temperature distribution observer for a cross-current battery and its application, provides a method for constructing a two-dimensional temperature distribution observer for a cross-current battery, which verifies and checks the measured values ​​of the inlet and outlet temperatures of the battery stack to predict the temperature distribution inside the battery. However, this method lacks the coupling process of electrochemistry and heat transfer, and it is difficult to directly verify the accuracy of the predicted value by simply starting from the measured values ​​of the inlet and outlet temperatures of the battery stack and matching the temperature inside the battery stack through linear or nonlinear calculations; The patent application with the publication number CN112599820A and the name of the prediction method of the quasi-three-dimensional multi-physical field coupled temperature distribution of the fuel cell stack discloses a prediction method of the quasi-three-dimensional multi-physical field coupled temperature distribution of the fuel cell stack, which obtains the relationship between the pressure drop and the flow rate and the stack flow distribution through the single cell pressure drop characteristics to calculate the stack temperature distribution and realize the prediction of the stack temperature distribution. However, this method ignores the influence of the complex mass transfer mechanism and geometric structure inside the stack, and also needs to tediously introduce other sub-models such as coolant, anode hydrogen, cathode air manifold, single cell flow field, etc. to establish the fluid network model at the stack level; The patent application with publication number CN115270633A, entitled Prediction method, system, device and medium for three-dimensional physical field of fuel cell, discloses a prediction method for three-dimensional physical field of fuel cell, which realizes online prediction of three-dimensional physical field inside fuel cell through several twin snapshots obtained based on offline simulation. However, this method requires a large amount of simulation data to train the twin model, and still does not solve the problem of excessive computing resources when simulating high-power stacks. Summary of the invention

[0004] The object of the present invention is to provide a fuel cell physical field simulation method, system and device to solve at least one of the above-mentioned technical problems existing in the prior art.

[0005] In the first aspect, in order to solve the above technical problems, the present invention provides a fuel cell physical field simulation method, including a temperature field simulation method, specifically comprising: Step a1, collecting temperature data of the battery stack; taking the ambient temperature as the initial temperature distribution; taking the temperature and heat flow of the battery stack cooling plate as boundary conditions; Step a2, constructing a PINN model (physical information neural network, a machine learning method that combines physical laws with neural networks): taking the three-dimensional coordinate value and time of the temperature field of the battery stack as input parameters; taking the temperature in the battery stack as the output result; defining the heat conduction formula of the battery stack and constructing a residual function; taking the initial temperature distribution and the boundary conditions as constraints and constructing a residual function; defining a loss function based on the multiple residual functions; Taking the temperature data as a training set, and training the PINN model for prediction by minimizing the loss function; The PINN model is evaluated by preset evaluation criteria, and the optimal model parameters are determined to obtain a trained PINN model; Step a3, reconstructing the temperature field through the trained PINN model, inputting the three-dimensional coordinate value and time of the temperature field of the battery stack, and outputting the corresponding temperature value; Through the above method, physical constraints such as the heat conduction formula of the fuel cell stack can be embedded into the concise PINN model. The trained PINN model can be used to predict the temperature field of the fuel cell, thereby avoiding complex model structure and massive simulation data, and realizing temperature field simulation with high efficiency and quality.

[0006] In a feasible embodiment, the specific method for collecting temperature data in step a1 is: placing a number of PCB flow field plates at a number of single cells in the battery stack, including a cathode PCB board and an anode PCB board, symmetrically attached to the two sides of the membrane electrode assembly (MEA); the PCB flow field plate is divided into a number of partitions, each partition is individually integrated with a temperature sensor and a resistor, which are used to respectively record the temperature and current density distribution data of the anode plate and the cathode plate in each partition.

[0007] In a feasible implementation, the specific method for collecting the temperature data in step a1 may also be: non-contact measurement of the surface temperature of the stack is performed by infrared thermal imaging equipment to obtain global data of the surface temperature field of the stack as the temperature data.

[0008] In a feasible implementation manner, the specific method for collecting the temperature data in step a1 may also be: embedding and arranging a plurality of optical fiber temperature sensors in the battery stack to collect temperature data, thereby improving the collection accuracy.

[0009] In a feasible implementation manner, the process of defining the heat conduction formula of the battery stack in step a2 includes: Step a21, substitute the heat generation formula and the heat dissipation formula into the heat conduction equation to obtain the heat conduction formula of the battery stack, which can be specifically expressed as: ; in, Indicates the effective density of the battery stack; Indicates the effective heat capacity of the battery stack; represents the effective thermal conductivity of the stack, Represents the heat source term of the battery stack, in units of ; Represents the heat dissipation flux of the battery stack.

[0010] Step a22, defining heat source items, including: ; Represents the electrochemical reaction heat, which mainly depends on the current density and working voltage at each moment. The specific calculation formula is: ; in, represents the difference in the local heat generation rate through the electrochemical reaction; It represents the equivalent current density of a single cell in amperes per square meter, which can be expressed as , that is, the battery stack Location The equivalent local current density at time ; Indicates the active area of ​​the battery in square meters; It represents the enthalpy change of the hydrogen-oxygen reaction, in joules per mole, and the specific value is about 286 kilojoules per mole; It represents the number of electrons transferred in the hydrogen-oxygen reaction, and its value is 2; Represents the Faraday constant, with a value of 96485 and a unit of ; Indicates that the battery stack is The working voltage at the moment, in volts, can be further expressed as , that is, the battery stack Location The local operating voltage at the moment; It represents ohmic heat loss, that is, the heat loss caused by the internal resistance of the battery stack. The specific calculation formula is: ; in, represents the difference in the rate of heat generation through ohmic losses, Indicates the equivalent internal resistance of the battery stack, in ohms; Indicates the concentration loss heat. Under low current density or good gas supply conditions, the concentration loss heat caused by the gas concentration difference can be ignored. In summary, the heat production formula can be derived as follows: ; Step a23, define the heat dissipation flux, including the convection heat dissipation formula with the outside world and the forced heat dissipation formula with the coolant: The convection heat dissipation formula is specifically: ; in, represents the heat dissipation by natural convection; Represents the convective heat transfer coefficient between the battery stack and the external environment, in units of ; Indicates the stack temperature in units of ; Indicates the external environment temperature in units of ; This is further described by Newton's law of cooling: ; in, represents the effective thermal conductivity of the stack, Represents the temperature distribution of the battery stack in the time and space domain; The forced heat dissipation formula assumes that the coolant evenly removes heat from the surface of the battery stack, specifically: ; in, Indicates the convective heat dissipation of the coolant; Indicates the convective heat transfer coefficient between the coolant and the plate; Indicates coolant temperature; This is further described by Newton's law of cooling: ; In summary, the heat dissipation formula can be derived as follows: ; in, Represents the heat dissipation flux of the battery stack, in units of ; The initial temperature of the battery stack is set to the external ambient temperature. The specific expression is: ; Set the coolant temperature to a constant value. The specific expression is: .

[0011] In a feasible implementation, the loss function in step a2 includes: partial differential equation (PDE) loss , Boundary loss , initial loss and data loss ; The residual function of the partial differential equation loss can be expressed as: ; in, represents the number of training points selected from the computational domain of the partial differential equation (i.e., the heat conduction equation); represents the residual of the partial differential equation; Represents the hyperparameters between neural network connection layers; represents the computational domain of the partial differential equation loss function; The residual function of the boundary loss can be expressed as: ; in, represents the number of training points selected from the computational domain of the cooling boundary (BC); represents the boundary conditions; Represents the computational domain of the boundary loss function; The residual function of the initial loss can be expressed as: ; in, represents the number of training points selected from the computational domain of the initial boundary (IC); represents the initial condition; Represents the computational domain of the initial loss function; The residual function of the data loss can be expressed as: ; in, Indicates the number of test points selected for the experiment; Residual function representing the real test data; Represents temperature sensor data; In this way, the PINN model can be trained in the subsequent order based on the temperature sensor data by minimizing the error between the output result (predicted temperature value) of the PINN model and the actual measured temperature value of the temperature sensor; Construct the overall loss set: ; Assign corresponding weights to each loss , construct the residual function of the overall loss, which can be specifically: ; in, Indicates overall loss; The loss functions represent partial differential equation loss, boundary loss, initial loss and data loss respectively; Represents the weight of each loss function.

[0012] In a feasible implementation, in step a2, an adaptive gradient optimization algorithm is used when training the PINN model to minimize the loss function.

[0013] In a feasible implementation, in step a2, an automatic differentiation tool is used when training the PINN model to calculate the derivatives of the temperature field with respect to space and time, thereby obtaining the physical constraint term (heat conduction formula of the battery stack).

[0014] In a feasible implementation manner, the preset evaluation criteria in step a2 include mean absolute error and determination coefficient; The specific calculation formula of the mean absolute error is: ; in, Indicates the number of test points; Indicates The actual measured temperature value of each test point; Indicates the predicted temperature value; The specific calculation formula of the determination coefficient is: ; in, represents the coefficient of determination; Indicates the average value of the actual measured temperature values.

[0015] In a feasible implementation manner, the step a3 further includes: when the temperature value exceeds a preset overheating threshold, generating overheating alarm information.

[0016] In the second aspect, based on the same inventive concept, the present application also provides a fuel cell physical field simulation system, including a temperature field simulation system, specifically including a data receiving module, a data processing module and a result generating module; The data receiving module is used to receive the temperature data of the battery stack, the ambient temperature data, the temperature and heat flow data of the battery stack cooling plate, and the three-dimensional coordinate value and time value of the battery stack temperature field; The data processing module includes a model unit, a training unit, an evaluation unit and a prediction unit; The model unit stores a PINN model: the input parameters are the three-dimensional coordinate values ​​and time values ​​of the temperature field of the battery stack; the output result is the temperature in the battery stack; the loss function includes a residual function and constraints; the residual function includes a heat conduction formula for the battery stack; the constraints include an initial temperature distribution and boundary conditions; The training unit uses the ambient temperature data as the initial temperature distribution, the temperature and heat flow data of the stack cooling plate as boundary conditions, and the temperature data as a training set, and trains the PINN model for prediction by minimizing a loss function; The evaluation unit evaluates the PINN model according to a preset evaluation standard to determine the optimal model parameters; The prediction unit calls the PINN model, inputs the three-dimensional coordinate value and time value of the temperature field of the fuel cell stack, and outputs the corresponding temperature value; The result generating module is used to send the temperature value externally.

[0017] On the third aspect, based on the same inventive concept, the present application also provides a fuel cell physical field simulation device, including a processor, a memory and a bus, wherein the memory stores instructions and data read by the processor, and the processor is used to call the instructions and data in the memory to execute the fuel cell physical field simulation method as described above, and the bus connects the functional components for transmitting information.

[0018] By adopting the above technical solution, the present invention has the following beneficial effects: A fuel cell physical field simulation method, system and device provided by the present invention can be combined with real-time sensor data, and by combining physical constraints and limited data points, accurate physical field distribution can still be obtained when data is insufficient; compared with traditional interpolation or extrapolation methods, this solution is more robust, and can provide accurate and dynamic predictions of physical fields such as temperature under complex geometric and operating conditions inside the fuel cell stack, which can help the fuel cell management system to monitor physical fields such as temperature in real time during operation, and can also provide early warning signals for local overheating, thereby avoiding fuel cell performance degradation or damage; compared with traditional CFD simulation methods, this solution avoids a large amount of grid division and iterative calculation, reduces calculation complexity, is more efficient in reconstructing the physical field, significantly reduces the amount of calculation, can achieve faster physical quantity prediction, and is suitable for online monitoring; this solution can ensure that high prediction accuracy and stability are maintained during long-term operation and different load changes, and avoid error accumulation; this solution can achieve high-density layout and multi-functional integration by integrating various sensors on a PCB board, reduces space occupancy, is easy to install, and basically does not expand the overall volume of the fuel cell stack. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0020] Figure 1 A flow chart of a method for simulating a temperature field of a fuel cell provided by an embodiment of the present invention; Figure 2 A diagram of a battery stack structure provided by an embodiment of the present invention; Figure 3 A schematic diagram of the arrangement of a PCB flow field plate provided in an embodiment of the present invention; Figure 4 A schematic diagram of the internal layout of a PCB flow field plate provided in an embodiment of the present invention; Figure 5 A diagram showing the definition process of the heat conduction formula for a battery stack provided in an embodiment of the present invention; Figure 6 A diagram of a temperature field simulation system for a fuel cell provided by an embodiment of the present invention; Reference numerals: 1-Membrane electrode assembly; 2-Cathode PCB board; 3-Anode PCB board; 4-Cathode bipolar plate; 5-Anode bipolar plate. DETAILED DESCRIPTION

[0021] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. 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.

[0022] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0023] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0024] The present invention is further explained below in conjunction with specific implementation modes.

[0025] It should also be noted that the following specific embodiments or specific implementations are a series of optimized settings listed in the present invention to further explain the specific content of the invention, and these settings can be used in combination or in association with each other.

[0026] Embodiment 1: like Figure 1 As shown, this embodiment provides a fuel cell physical field simulation method, including a temperature field simulation method, specifically including: Step a1, collecting temperature data of the battery stack; taking the ambient temperature as the initial temperature distribution; taking the temperature and heat flow of the battery stack cooling plate as boundary conditions; Step a2, constructing a PINN model: taking the three-dimensional coordinate value and time of the temperature field of the battery stack as input parameters; taking the temperature in the battery stack as the output result; defining the heat conduction formula of the battery stack and constructing a residual function; taking the initial temperature distribution and the boundary conditions as constraints; and defining a loss function based on the residual function and the constraints; Taking the temperature data as a training set, and training the PINN model for prediction by minimizing the loss function; The PINN model is evaluated by preset evaluation criteria, and the optimal model parameters are determined to obtain a trained PINN model; Step a3: reconstruct the temperature field through the trained PINN model, input the three-dimensional coordinate value and time of the temperature field of the battery stack, and output the corresponding temperature value.

[0027] Furthermore, the specific method for collecting temperature data in step a1 is as follows: Figure 2-3As shown, in the battery stack, for every ten single cells, a group of PCB flow field plates are embedded in one single cell, including a cathode PCB board 2 and an anode PCB board 3, which are symmetrically attached to the two sides of the membrane electrode assembly 1, that is, the cathode PCB board 2 is embedded between the mold electrode assembly 1 of the single cell and the cathode bipolar plate 4, and the anode PCB board 3 is embedded between the mold electrode assembly 1 of the single cell and the anode bipolar plate 5; 150 square partitions are divided on the PCB flow field plate, such as Figure 4 As shown, there are 15 columns horizontally and 10 rows vertically, and each partition is individually integrated with a temperature sensor and a resistor to respectively record the temperature and current density distribution data of the anode plate and the cathode plate in each partition.

[0028] Furthermore, if Figure 5 As shown, the process of defining the heat conduction formula of the battery stack in step a2 includes: Step a21, substitute the heat generation formula and the heat dissipation formula into the heat conduction equation to obtain the heat conduction formula of the battery stack, which can be specifically expressed as: ; in, Indicates the effective density of the battery stack; Indicates the effective heat capacity of the battery stack; represents the effective thermal conductivity of the stack, Represents the heat source term of the battery stack, in units of ; Represents the heat dissipation flux of the battery stack.

[0029] Step a22, defining heat source items, including: ; Represents the electrochemical reaction heat, which mainly depends on the current density and working voltage at each moment. The specific calculation formula is: ; in, represents the difference in the local heat generation rate through the electrochemical reaction, It represents the equivalent current density of a single cell in amperes per square meter, which can be expressed as , that is, the battery stack Location The equivalent local current density at time ; Indicates the active area of ​​the battery in square meters; It represents the enthalpy change of the hydrogen-oxygen reaction, in joules per mole, and the specific value is about 286 kilojoules per mole; It represents the number of electrons transferred in the hydrogen-oxygen reaction, and its value is 2; Represents the Faraday constant, with a value of 96485 and a unit of ; Indicates that the battery stack is The working voltage at the moment, in volts, can be further expressed as , that is, the battery stack Location The local operating voltage at the moment; It represents ohmic heat loss, that is, the heat loss caused by the internal resistance of the battery stack. The specific calculation formula is: ; in, represents the difference in the rate of heat generation through ohmic losses, Indicates the equivalent internal resistance of the battery stack, in ohms; Indicates the concentration loss heat. Under low current density or good gas supply conditions, the concentration loss heat caused by the gas concentration difference can be ignored. In summary, the heat production formula can be derived as follows: ; Step a23, define the heat dissipation flux, including the convection heat dissipation formula with the outside world and the forced heat dissipation formula with the coolant: The convection heat dissipation formula is specifically: ; in, represents the heat dissipation by natural convection; Represents the convective heat transfer coefficient between the battery stack and the external environment, in units of ; Indicates the stack temperature in units of ; Indicates the external environment temperature in units of ; This is further described by Newton's law of cooling: ; in, represents the effective thermal conductivity of the stack, Represents the temperature distribution of the battery stack in the time and space domain; The forced heat dissipation formula assumes that the coolant evenly removes heat from the surface of the battery stack, specifically: ; in, Indicates the convective heat dissipation of the coolant; Indicates the convective heat transfer coefficient between the coolant and the plate; Indicates coolant temperature; This is further described by Newton's law of cooling: ; In summary, the heat dissipation formula can be derived as follows: ; in, Indicates the heat dissipation of the battery stack, in units of ; The initial temperature of the battery stack is set to the external ambient temperature. The specific expression is: ; Set the coolant temperature to a constant value. The specific expression is: ; In a feasible implementation, the loss function in step a2 includes: partial differential equation (PDE) loss , Boundary loss , initial loss and data loss ; The residual function of the partial differential equation loss can be expressed as: ; in, represents the number of training points selected from the computational domain of the partial differential equation (heat conduction equation); represents the residual of the partial differential equation; Represents the hyperparameters between neural network connection layers; represents the computational domain of the partial differential equation loss function; The residual function of the boundary loss can be expressed as: ; in, represents the number of training points selected from the computational domain of the cooling boundary (BC); represents the boundary conditions; Represents the computational domain of the boundary loss function; The residual function of the initial loss can be expressed as: ; in, represents the number of training points selected from the computational domain of the initial boundary (IC); represents the initial condition; Represents the computational domain of the initial loss function; The residual function of the data loss can be expressed as: ; in, Indicates the number of test points selected for the experiment; Residual function representing the real test data; Represents temperature sensor data; Construct the overall loss set: ; Assign corresponding weights to each loss , construct the residual function of the overall loss, which can be specifically: ; in, Indicates overall loss; The loss functions represent partial differential equation loss, boundary loss, initial loss and data loss respectively; Represents the weight of each loss function.

[0030] Furthermore, in step a2, an adaptive gradient optimization algorithm (such as the Adam algorithm) is used when training the PINN model to minimize the loss function.

[0031] Furthermore, in step a2, an automatic differentiation tool (such as autograd in PyTorch) is used when training the PINN model to calculate the derivatives of the temperature field with respect to space and time, thereby obtaining the physical constraint term (heat conduction equation of the battery stack).

[0032] Furthermore, the preset evaluation criteria in step a2 include mean absolute error and coefficient of determination; The specific calculation formula of the mean absolute error is: ; in, Indicates the number of test points; Indicates The actual measured temperature value of each test point; Indicates the predicted temperature value; The specific calculation formula of the determination coefficient is: ; in, represents the coefficient of determination; Indicates the average value of the actual measured temperature values.

[0033] In this way, the PINN model can be trained in the subsequent sequence based on the temperature sensor data by minimizing the error between the output result (predicted temperature value) of the PINN model and the actual measured temperature value of the temperature sensor.

[0034] Furthermore, in step a2, an adaptive gradient optimization algorithm is used when training the PINN model to minimize the loss function.

[0035] Furthermore, in step a2, an automatic differentiation tool is used when training the PINN model to calculate the derivatives of the temperature field with respect to space and time, thereby obtaining the physical constraint term (heat conduction formula of the battery stack).

[0036] Embodiment 2: like Figure 6 As shown, this embodiment provides a fuel cell physical field simulation system, including a temperature field simulation system, specifically including a data receiving module, a data processing module and a result generating module; The data receiving module is used to receive the temperature data of the battery stack, the ambient temperature data, the temperature and heat flow data of the battery stack cooling plate, and the three-dimensional coordinate value and time value of the battery stack temperature field; The data processing module includes a model unit, a training unit, an evaluation unit and a prediction unit; The model unit stores a PINN model: the input parameters are the three-dimensional coordinate values ​​and time values ​​of the temperature field of the battery stack; the output result is the temperature in the battery stack; the loss function includes a residual function and constraints; the residual function includes a heat conduction formula for the battery stack; the constraints include an initial temperature distribution and boundary conditions; The training unit uses the ambient temperature data as the initial temperature distribution, the temperature and heat flow data of the stack cooling plate as boundary conditions, and the temperature data as a training set, and trains the PINN model for prediction by minimizing a loss function; The evaluation unit evaluates the PINN model according to a preset evaluation standard to determine the optimal model parameters; The prediction unit calls the PINN model, inputs the three-dimensional coordinate value and time value of the temperature field of the fuel cell stack, and outputs the corresponding temperature value; The result generating module is used to send the temperature value externally.

[0037] Embodiment three: This embodiment provides a fuel cell physical field simulation device, including a processor, a memory and a bus, wherein the memory stores instructions and data read by the processor, the processor is used to call the instructions and data in the memory to execute the fuel cell physical field simulation method as described above, and the bus connects the functional components for transmitting information.

[0038] In another embodiment, the present solution can be implemented by an integrated device, which may include corresponding modules for performing each or several steps in the above-mentioned embodiments. The module may be one or more hardware modules specially configured to perform the corresponding steps, or implemented by a processor configured to perform the corresponding steps, or stored in a computer-readable medium for implementation by a processor, or implemented by some combination.

[0039] The processor performs the various methods and processes described above. For example, the method implementation in the present solution can be implemented as a software program, which is tangibly contained in a machine-readable medium, such as a memory. In some embodiments, part or all of the software program can be loaded and / or installed via a memory and / or a communication interface. When the software program is loaded into the memory and executed by the processor, one or more steps in the method described above can be executed. Alternatively, in other embodiments, the processor can be configured to perform one of the above methods by any other appropriate means (for example, by means of firmware).

[0040] The device can be implemented using a bus architecture. The bus architecture can include any number of interconnecting buses and bridges, depending on the specific application and overall design constraints of the hardware. The bus connects various circuits including one or more processors, memories, and / or hardware modules together. The bus can also connect various other circuits such as peripherals, voltage regulators, power management circuits, external antennas, etc.

[0041] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Component (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc.

[0042] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A fuel cell physical field simulation method, characterized in that: Including temperature field simulation methods, specifically including: Step a1, collecting temperature data of the battery stack; taking the ambient temperature as the initial temperature distribution; taking the temperature and heat flow of the battery stack cooling plate as boundary conditions; Step a2, constructing a PINN model: taking the three-dimensional coordinate value and time of the temperature field of the battery stack as input parameters; taking the temperature in the battery stack as the output result; defining the heat conduction formula of the battery stack and constructing a residual function; taking the initial temperature distribution and the boundary conditions as constraints; and defining a loss function based on the residual function and the constraints; Taking the temperature data as a training set, and training the PINN model for prediction by minimizing the loss function; The PINN model is evaluated by preset evaluation criteria, and the optimal model parameters are determined to obtain a trained PINN model; Step a3: reconstruct the temperature field through the trained PINN model, input the three-dimensional coordinate value and time of the temperature field of the battery stack, and output the corresponding temperature value.

2. The method according to claim 1, characterized in that The specific method for collecting the temperature data in step a1 is as follows: placing a number of PCB flow field plates at a number of cells in the battery stack, including a cathode PCB board and an anode PCB board, symmetrically attached to the two sides of the membrane electrode assembly; The PCB flow field plate is divided into several partitions, each of which is individually integrated with a temperature sensor and a resistor for respectively recording the temperature and current density distribution data of the anode plate and the cathode plate in each partition.

3. The method according to claim 1, characterized in that The heat conduction formula of the battery stack in step a2 is specifically expressed as: ； in, Indicates the effective density of the battery stack; Indicates the effective heat capacity of the battery stack; Indicates the effective thermal conductivity of the battery stack; Indicates the stack temperature; Indicates the moment; It represents the heat source term of the battery stack, and the specific formula is: ； in, represents the difference in the local heat generation rate through the electrochemical reaction; Represents the equivalent current density of a single cell; Indicates the active area of ​​the battery; represents the enthalpy change of the hydrogen-oxygen reaction; It represents the number of electrons transferred in the hydrogen-oxygen reaction; represents the Faraday constant; Indicates that the battery stack is Working voltage at all times; represents the difference in the rate of heat production through ohmic losses; Indicates the equivalent internal resistance of the battery stack; It represents the heat dissipation flux of the battery stack. The specific formula is: ； in, Indicates the convective heat transfer coefficient between the coolant and the plate; Indicates coolant temperature; Indicates the convective heat transfer coefficient between the battery stack and the external environment; Indicates the external environment temperature.

4. The method according to claim 3, characterized in that: The loss function in step a2 includes: partial differential equation loss , Boundary loss , initial loss and data loss ; The residual function of the partial differential equation loss is expressed as: ； in, represents the number of training points selected from the computational domain of the partial differential equation; represents the residual of the partial differential equation; Represents the hyperparameters between neural network connection layers; represents the computational domain of the partial differential equation loss function; The residual function of the boundary loss is expressed as: ； in, represents the number of training points selected from the computational domain of the heat dissipation boundary; represents the boundary conditions; Represents the computational domain of the boundary loss function; The residual function of the initial loss is expressed as: ； in, represents the number of training points selected from the computational domain of the initial boundary; represents the initial condition; Represents the computational domain of the initial loss function; The residual function of the data loss can be expressed as: ； in, Indicates the number of test points selected for the experiment; Residual function representing the real test data; Represents temperature sensor data; Construct the overall loss set: ; Assign corresponding weights to each loss , construct the residual function of the overall loss, specifically: ； in, Indicates overall loss; The loss functions represent partial differential equation loss, boundary loss, initial loss and data loss respectively; Represents the corresponding weight of each loss function.

5. The method according to claim 3, characterized in that: The value is 286 kilojoules per mole.

6. The method according to claim 3, characterized in that The value is 2.

7. The method according to claim 3, characterized in that The value is 96485, and the unit is .

8. The method according to claim 1, characterized in that: The preset evaluation criteria in step a2 include mean absolute error and coefficient of determination; The specific calculation formula of the mean absolute error is: ； in, Indicates the number of test points; Indicates The actual measured temperature value of each test point; Indicates the predicted temperature value; The specific calculation formula of the determination coefficient is: ； in, represents the coefficient of determination; Indicates the average value of the actual measured temperature values.

9. A fuel cell physical field simulation system, characterized in that: It includes a temperature field simulation system, which specifically includes a data receiving module, a data processing module and a result generating module; The data receiving module is used to receive the temperature data of the battery stack, the ambient temperature data, the temperature and heat flow data of the battery stack cooling plate, and the three-dimensional coordinate value and time value of the battery stack temperature field; The data processing module includes a model unit, a training unit, an evaluation unit and a prediction unit; The model unit stores a PINN model: the input parameters are the three-dimensional coordinate value and time value of the temperature field of the battery stack; the output result is the temperature in the battery stack; The loss function includes a residual function and constraints; the residual function includes a heat conduction formula for the battery stack; the constraints include an initial temperature distribution and boundary conditions; The training unit uses the ambient temperature data as the initial temperature distribution, the temperature and heat flow data of the stack cooling plate as boundary conditions, and the temperature data as a training set, and trains the PINN model for prediction by minimizing a loss function; The evaluation unit evaluates the PINN model according to a preset evaluation standard to determine the optimal model parameters; The prediction unit calls the PINN model, inputs the three-dimensional coordinate value and time value of the temperature field of the fuel cell stack, and outputs the corresponding temperature value; The result generating module is used to send the temperature value externally.

10. A fuel cell physical field simulation device, characterized in that: It includes a processor, a memory and a bus, wherein the memory stores instructions and data read by the processor, the processor is used to call the instructions and data in the memory to execute any method as claimed in claim 1-8, and the bus connects the functional components for transmitting information.

Images