Fuel cell internal water state monitoring method, system and management method

Abstract

The invention discloses a fuel cell internal water state monitoring method and system and a management method. The method comprises the following steps: training a fuel cell water state observation model by using a training sample set to obtain training prediction data; calculating a training loss value between the training prediction data and the real label according to component diffusion, liquid water transmission, modal water transmission, ion and electron transmission and heat transmission in a direction perpendicular to a proton exchange membrane in the fuel cell; optimizing parameters of the fuel cell water state observation model according to the training loss value to obtain a trained fuel cell water state observation model; inputting the real-time operation data into a pre-constructed fuel cell water state observation model to obtain evaluation data of the water content in the fuel cell; the water state in the fuel cell can be quickly responded and accurately monitored, and the fuel cell is regulated and controlled according to the water state in the fuel cell, so that a membrane dry fault and a water logging fault of the fuel cell are avoided.

Description

Technical Field

[0001] The present invention belongs to the technical field of fuel cell health management, and specifically relates to a method, system and management method for monitoring the internal water status of a fuel cell. Background Art

[0002] In the electrification process of construction machinery, medium and large construction machinery are more suitable for the development of fuel cell routes. They have the characteristics of high efficiency, low noise and cleanliness, and have advantages in the fields of mining dump trucks, tractors, sanitation vehicles and maintenance vehicles. Construction machinery generally has the characteristics of large load and fast load change. When the vehicle fuel cell system is working, the working conditions will change with the rapid changes of the vehicle working conditions, and the internal water content of the fuel cell will significantly affect the output performance.

[0003] When the fuel cell is working, the electrochemical reaction will produce a large amount of water, and some water will also be carried when the fuel and air are supplied. If there is too much water inside the battery, it may block the gas flow channel, accumulate on the surface of the gas diffusion layer to hinder the transport of reactants, or cover the surface of the catalyst to reduce the effective reaction area, resulting in water flooding failure, reducing the efficiency and maximum output power of the fuel cell. Working for a long time under a high current load may reduce the water content of the proton exchange membrane, reduce the proton passage rate, form a membrane dry failure, cause local hot spots in the proton exchange membrane, and cause permanent damage. Therefore, rapid response and accurate monitoring of the water status inside the fuel cell are of great significance to ensure the efficient and stable operation of the fuel cell. Summary of the invention

[0004] The present invention provides a method, system and management method for monitoring the internal water status of a fuel cell, which can quickly respond to and accurately monitor the internal water status of a fuel cell, and regulate the fuel cell according to the internal water status of the fuel cell to avoid membrane dry failure and water flooding failure of the fuel cell.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] A first aspect of the present invention provides a method for monitoring the internal water status of a fuel cell, comprising:

[0007] Monitor the fuel cell to obtain real-time operation data; input the real-time operation data into a pre-built fuel cell water state observation model to obtain water content assessment data in the fuel cell;

[0008] The construction process of the fuel cell water state observation model includes:

[0009] Obtaining the running storage data of the fuel cell, preprocessing the running storage data and adding real labels to obtain a training sample set;

[0010] A fuel cell water state observation model is constructed based on a neural convolutional network, and the fuel cell water state observation model is trained using a training sample set to obtain training prediction data; the training loss value between the training prediction data and the true label is calculated based on component diffusion, liquid water transport, modal water transport, ion and electron transport, and heat transport in the direction perpendicular to the proton exchange membrane in the fuel cell; the Adam optimizer is used to optimize the parameters of the fuel cell water state observation model according to the training loss value, and the iterative training process is repeated until the set training cycle is reached to output the trained fuel cell water state observation model.

[0011] Furthermore, the real-time operation data and operation storage data of the fuel cell include stack voltage, stack current, stack inlet air pressure, stack outlet air pressure, stack inlet hydrogen pressure, stack outlet hydrogen pressure, stack inlet coolant temperature, stack outlet coolant temperature and stack inlet air flow.

[0012] Furthermore, the process of preprocessing the running storage data includes:

[0013] Check the missing values ​​in the running stored data and fill them in through the interpolation algorithm;

[0014] The outliers in the running storage data are identified by the Z-score (standard score) algorithm, and the outliers are replaced by the mean or median in the running storage data; then the running storage data are normalized and standardized.

[0015] Furthermore, the training loss value between the training prediction data and the true label is calculated according to the component diffusion, liquid water transport, modal water transport, ion and electron transport and heat transport in the direction perpendicular to the proton exchange membrane in the fuel cell, and the process includes:

[0016] The fuel cell comprises anode and cathode bipolar plates, an anode flow channel, a cathode flow channel, an anode diffusion layer, a cathode diffusion layer, an anode microporous layer, a cathode microporous layer, a cathode catalyst layer, an anode catalyst layer and a proton exchange membrane;

[0017] Calculate the data term loss between the training prediction data and the true label, expressed as: ;

[0018] In the formula, is the data item loss function; Predict data items in the training data; is the data item in the real label; is the square sum function of absolute values;

[0019] Calculate the boundary loss between the training prediction data and the true label, expressed as: ;

[0020] In the formula, is the boundary loss, For training prediction data Gas composition at the interface; For training prediction data Liquid water components at the interface; For training prediction data Modal water component at the interface; The interface is the interface between the anode catalyst layer and the proton exchange membrane and the interface between the cathode catalyst layer and the proton exchange membrane; The interface is the interface between the anode catalyst layer and the anode microporous layer and the interface between the cathode catalyst layer and the cathode microporous layer; is the square sum function of absolute values; For training prediction data Liquid pressure at the boundary; For training prediction data Gas phase pressure at the boundary; For training prediction data Capillary pressure on the boundary; The boundaries are the interface between the anode flow channel and the anode diffusion layer and the interface between the cathode flow channel and the cathode diffusion layer.

[0021] Calculate physical information losses based on component transport, liquid water transport, modal water transport, ion and electron transport, and heat transport in the direction perpendicular to the proton exchange membrane in the fuel cell;

[0022] The training loss value is calculated based on the data item loss, boundary loss and physical information loss, and the expression formula is: ;

[0023] In the formula, is the training loss value; is the physical information loss function; , , is the weight of each loss item.

[0024] Furthermore, the physical information loss calculated based on the component diffusion, liquid water transport, modal water transport, ion and electron transport, and heat transport in the direction perpendicular to the proton exchange membrane in the fuel cell is expressed as follows: ;

[0025] In the formula, is the diffusion state of the components; It is a liquid water transmission state; is the modal water transport state; It is the ion transport state; It is the state of electronic transmission; It is the state of heat transfer;

[0026] The expression formula for calculating the diffusion state of the components is: ;

[0027] In the formula, For time; are components, and the components are oxygen, hydrogen, nitrogen and water vapor; is the porosity; is the liquid water saturation; For each component The molar concentration of is the dimension perpendicular to the proton exchange membrane; For components The effective diffusion coefficient of For components The universal source term of ;

[0028] The expression formula for calculating the liquid water transmission state is: ;

[0029] In the formula, is the density of liquid water; is the intrinsic permeability of the porous medium; is the dynamic viscosity of liquid water; Source term generated for liquid water;

[0030] The expression formula for calculating the modal water transmission state is: ;

[0031] In the formula, is the dry film density; is the dry film equivalent; is the volume fraction of ionomer in the catalyst layer; is the modal water content; is the effective diffusion coefficient of modal water; Source terms generated for modal water;

[0032] The expression formula for calculating the ion transport state is: ;

[0033] In the formula, is the ionic conductivity of the proton exchange membrane; is the ionic potential; is the source term of the ionic potential;

[0034] The expression formula for calculating the electron transfer state is: ;

[0035] In the formula, is the intrinsic conductivity of the porous electrode skeleton; is the electron potential; is the source term of the electron potential;

[0036] The expression formula for calculating the heat transfer state is: ;

[0037] In the formula, is the equivalent heat capacity; is the equivalent thermal conductivity; is the temperature; is the general source term in the energy equation.

[0038] A second aspect of the present invention provides a fuel cell internal water status monitoring system, comprising:

[0039] A monitoring unit is used to monitor the fuel cell to obtain real-time operation data; the real-time operation data is input into a pre-built fuel cell water state observation model to obtain water content evaluation data in the fuel cell;

[0040] An acquisition unit, used to acquire the operation storage data of the fuel cell, pre-process the operation storage data and add real labels to obtain a training sample set;

[0041] Model building unit, building a fuel cell water state observation model based on a neural convolutional network;

[0042] The training unit is used to train the fuel cell water state observation model using the training sample set to obtain training prediction data; calculate the training loss value between the training prediction data and the true label according to the component diffusion, liquid water transmission, modal water transmission, ion and electron transmission and heat transmission in the direction perpendicular to the proton exchange membrane in the fuel cell; use the Adam optimizer to optimize the parameters of the fuel cell water state observation model according to the training loss value, repeat the iterative training process until the set training cycle is reached to output the trained fuel cell water state observation model.

[0043] Furthermore, the monitoring unit is arranged in the fuel cell controller; the acquisition unit, the model building unit and the training unit are arranged in the cloud service platform; the fuel cell controller is configured with a data storage module, a terminal communication box and an OTA module; the fuel cell controller exchanges information through the terminal communication box; the OTA module is used to update the parameters of the fuel cell water state observation model; the data storage module is used to store the real-time operation data of the fuel cell and the water content evaluation data in the fuel cell.

[0044] Furthermore, the real-time operation data and operation storage data of the fuel cell include stack voltage, stack current, stack air pressure, stack air pressure, stack hydrogen pressure, stack hydrogen pressure, stack coolant temperature, stack coolant temperature and stack air flow;

[0045] The monitoring unit is connected to a sensor group via a DC / DC converter; the sensor group includes a first pressure sensor for monitoring the air pressure entering the stack, a second pressure sensor for monitoring the air pressure leaving the stack, a third pressure sensor for monitoring the hydrogen pressure entering the stack, a fourth pressure sensor for monitoring the hydrogen pressure leaving the stack, a first temperature sensor for monitoring the coolant temperature entering the stack, a second temperature sensor for monitoring the coolant temperature leaving the stack, and a flow sensor for monitoring the air flow entering the stack.

[0046] The third aspect of the present invention provides a fuel cell management method, comprising: using the fuel cell internal water status monitoring method described in the first aspect to monitor the fuel motor to obtain the water content evaluation data in the fuel cell; the water content evaluation data in the fuel cell includes liquid water saturation and modal water content ;

[0047] When the modal water content of the anode in the fuel cell When the anode side membrane of the fuel cell is dry, it is determined that the anode side membrane of the fuel cell is dry. is the membrane dry threshold; controls the drainage frequency and exhaust frequency of the fuel cell minus the set value;

[0048] When the modal water content of the cathode in the fuel cell When the cathode side membrane of the fuel cell is dry, the cathode air flow rate of the fuel cell is controlled to be less than the set value;

[0049] When the liquid water saturation of the anode in the fuel cell When the anode side of the fuel cell is flooded, it is determined that the fuel cell has a flooding phenomenon; wherein, The membrane dry threshold is used to control the drainage frequency and exhaust frequency of the fuel cell to increase the set value;

[0050] When the liquid water saturation of the cathode in the fuel cell When the fuel cell is flooded on the cathode side, the fuel cell is judged to have flooding; the cathode air intake flow of the fuel cell is controlled to increase by the set value.

[0051] Furthermore, when the modal water content in the anode or cathode of the fuel cell is The time is greater than the limit , controlling the operating temperature of the fuel cell to drop to a set minimum value through a cooling device in the fuel cell;

[0052] When the liquid water saturation in the anode or cathode of the fuel cell The time is greater than the limit , the operating temperature of the fuel cell is controlled to be lowered to the set maximum value through the cooling device in the fuel cell.

[0053] The fourth aspect of the present invention provides an electronic device, comprising a storage medium and a processor; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the method for monitoring the internal water status of a fuel cell described in the first aspect.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] The present invention constructs a fuel cell water state observation model based on a neural convolutional network, and uses a training sample set to train the fuel cell water state observation model; in the process of calculating the training loss value, physical processes such as component diffusion, liquid water transmission, modal water transmission, ion and electron transmission, and heat transmission in the direction perpendicular to the proton exchange membrane are considered, which helps the model learn deeper physical laws and enhance the generalization ability of the model; accurately calculates the loss value and optimizes the model parameters to improve the model's prediction accuracy for the fuel cell water state.

[0056] The monitoring unit of the present invention is used to monitor the fuel cell to obtain real-time operation data; the real-time operation data is input into a pre-constructed fuel cell water state observation model to obtain water content evaluation data in the fuel cell; the monitoring unit can predict the water state of the fuel cell in real time, provide decision support for water management control, and adjust operating parameters in time to avoid flooding or membrane drying failures; through accurate water state prediction and management, the damage to the fuel cell caused by improper water management is reduced and its service life is extended. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 A flow chart of a method for monitoring the internal water status of a fuel cell provided in Example 1 of the present invention;

[0058] Figure 2 A structural diagram of a fuel cell and a sensor group provided in Example 2 of the present invention;

[0059] Figure 3 This is a flow chart of a fuel cell management method provided in Example 3 of the present invention.

[0060] In the figure: 1. First pressure sensor, 2. Second pressure sensor, 3. Third pressure sensor, 4. Fourth pressure sensor, 5. First temperature sensor, 6. Second temperature sensor, 7. Flow sensor DETAILED DESCRIPTION

[0061] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.

[0062] Example 1

[0063] like Figure 1 As shown, this embodiment provides an intelligent management method for a fuel cell, comprising:

[0064] The construction process of the fuel cell water state observation model includes:

[0065] The operation and storage data of the fuel cell are obtained, wherein the real-time operation data and the operation and storage data of the fuel cell include the stack voltage, the stack current, the stack inlet air pressure, the stack outlet air pressure, the stack inlet hydrogen pressure, the stack outlet hydrogen pressure, the stack inlet coolant temperature, the stack outlet coolant temperature and the stack inlet air flow rate.

[0066] The process of preprocessing running storage data includes:

[0067] Check the missing values ​​in the running stored data and fill them in through the interpolation algorithm;

[0068] The outliers in the running storage data are identified by the Z-score (standard score) algorithm, and the outliers are replaced by the mean or median in the running storage data; then the running storage data are normalized and standardized.

[0069] After preprocessing the running storage data, adding real labels to obtain a training sample set; constructing a fuel cell water state observation model based on a neural convolutional network, the fuel cell water state observation model includes 8 hidden layers, each hidden layer has 64 neuron nodes, and uses a hierarchical adaptive tanh activation function; using the training sample set to train the fuel cell water state observation model to obtain training prediction data;

[0070] The training loss between the training prediction data and the true label is calculated based on the component diffusion, liquid water transport, modal water transport, ion and electron transport, and heat transport in the direction perpendicular to the proton exchange membrane in the fuel cell. The process includes:

[0071] The fuel cell comprises anode and cathode bipolar plates, an anode flow channel, a cathode flow channel, an anode diffusion layer, a cathode diffusion layer, an anode microporous layer, a cathode microporous layer, a cathode catalyst layer, an anode catalyst layer and a proton exchange membrane;

[0072] Calculate the data term loss between the training prediction data and the true label, expressed as: ;

[0073] In the formula, is the data item loss function; Predict data items in the training data; is the data item in the real label; is the square sum function of absolute values;

[0074] Calculate the boundary loss between the training prediction data and the true label, expressed as: ;

[0075] In the formula, is the boundary loss, For training prediction data Gas composition at the interface; For training prediction data Liquid water components at the interface; For training prediction data Modal water component at the interface; The interface is the interface between the anode catalyst layer and the proton exchange membrane and the interface between the cathode catalyst layer and the proton exchange membrane; The interface is the interface between the anode catalyst layer and the anode microporous layer and the interface between the cathode catalyst layer and the cathode microporous layer; is the square sum function of absolute values; For training prediction data Liquid pressure at the boundary; For training prediction data Gas phase pressure at the boundary; For training prediction data Capillary pressure on the boundary; The boundaries are the interface between the anode flow channel and the anode diffusion layer and the interface between the cathode flow channel and the cathode diffusion layer.

[0076] The physical information loss calculated based on component diffusion, liquid water transport, modal water transport, ion and electron transport, and heat transport in the direction perpendicular to the proton exchange membrane in the fuel cell is expressed as: ;

[0077] In the formula, is the diffusion state of the components; It is a liquid water transmission state; is the modal water transport state; It is the ion transport state; It is the state of electronic transmission; It is the state of heat transfer;

[0078] The expression formula for calculating the diffusion state of the components is: ;

[0079] In the formula, For time; are components, and the components are oxygen, hydrogen, nitrogen and water vapor; is the porosity; is the liquid water saturation; For each component The molar concentration of is the dimension perpendicular to the proton exchange membrane; For components The effective diffusion coefficient of For components The universal source term of , which is generated by electrochemical reactions or gas-liquid phase changes;

[0080] The expression formula for calculating the liquid water transmission state is: ;

[0081] In the formula, is the density of liquid water; is the intrinsic permeability of the porous medium; is the dynamic viscosity of liquid water; The source term for liquid water generation, which is affected by evaporation / condensation or membrane water absorption / release processes;

[0082] The expression formula for calculating the modal water transmission state is: ;

[0083] In the formula, is the dry film density; is the dry film equivalent; is the volume fraction of ionomer in the catalyst layer; is the modal water content; is the effective diffusion coefficient of modal water, which is usually related to temperature and modal water content; The source term generated for the modal water is affected by the nonequilibrium adsorption / desorption process between the modal water and water vapor and the electroosmotic drag process;

[0084] The expression formula for calculating the ion transport state is: ;

[0085] In the formula, is the ionic conductivity of the proton exchange membrane; is the ionic potential; is the source term of the ionic potential;

[0086] The expression formula for calculating the electron transfer state is: ;

[0087] In the formula, is the intrinsic conductivity of the porous electrode skeleton; is the electron potential; is the source term of the electron potential;

[0088] The expression formula for calculating the heat transfer state is: ;

[0089] In the formula, is the equivalent heat capacity; is the equivalent thermal conductivity; is the temperature; is the general source term in the energy equation.

[0090] The training loss value is calculated based on the data item loss, boundary loss and physical information loss, and the expression formula is: ;

[0091] In the formula, is the training loss value; is the physical information loss function; , , is the weight of each loss item.

[0092] In this embodiment, physical processes such as component diffusion, liquid water transport, modal water transport, ion and electron transport, and heat transport in the direction perpendicular to the proton exchange membrane are considered during the calculation of the training loss value, which helps the model learn deeper physical laws and enhance the generalization ability of the model; accurately calculate the loss value and optimize the model parameters to improve the model's prediction accuracy for the water state of the fuel cell.

[0093] The Adam optimizer is used to optimize the parameters of the fuel cell water state observation model according to the training loss value, and the iterative training process is repeated until the set training cycle is reached to output the trained fuel cell water state observation model; in this embodiment, the training cycle is 20, and the initial learning rate is 0.001. Through the neural convolution network and optimization algorithm, the model can learn complex nonlinear relationships and improve the prediction accuracy of the fuel cell water state; the trained model can adapt to different operating conditions and environmental changes, and enhance the robustness of the system.

[0094] Monitor the fuel cell to obtain real-time operating data; input the real-time operating data into a pre-built fuel cell water state observation model to obtain water content assessment data in the fuel cell; this embodiment can predict the water state of the fuel cell in real time, provide decision support for water management control, and adjust operating parameters in time to avoid flooding or membrane dry failure; through accurate water state prediction and management, reduce fuel cell damage caused by improper water management and extend its service life.

[0095] Example 2

[0096] This embodiment provides a fuel cell internal water state monitoring system, the fuel cell internal water state monitoring system is used to execute the fuel cell internal water state monitoring method described in Example 1, and the fuel cell internal water state monitoring system includes:

[0097] A monitoring unit is used to monitor the fuel cell to obtain real-time operation data; the real-time operation data is input into a pre-built fuel cell water state observation model to obtain water content evaluation data in the fuel cell;

[0098] An acquisition unit, used to acquire the operation storage data of the fuel cell, pre-process the operation storage data and add real labels to obtain a training sample set;

[0099] Model building unit, building a fuel cell water state observation model based on a neural convolutional network;

[0100] The training unit is used to train the fuel cell water state observation model using the training sample set to obtain training prediction data; calculate the training loss value between the training prediction data and the true label according to the component diffusion, liquid water transmission, modal water transmission, ion and electron transmission and heat transmission in the direction perpendicular to the proton exchange membrane in the fuel cell; use the Adam optimizer to optimize the parameters of the fuel cell water state observation model according to the training loss value, repeat the iterative training process until the set training cycle is reached to output the trained fuel cell water state observation model.

[0101] The monitoring unit is arranged in the fuel cell controller; the acquisition unit, model building unit and training unit are arranged in the cloud service platform; the fuel cell controller is configured with a data storage module, a terminal communication box and an OTA module; the fuel cell controller exchanges information through the terminal communication box; the OTA module is used to update the parameters of the fuel cell water state observation model; the data storage module is used to store the real-time operation data of the fuel cell and the water content evaluation data in the fuel cell.

[0102] The real-time operation data and operation storage data of the fuel cell include stack voltage, stack current, stack air pressure, stack air pressure, stack hydrogen pressure, stack hydrogen pressure, stack coolant temperature, stack coolant temperature and stack air flow;

[0103] The monitoring unit is connected to a sensor group via a DC / DC converter; the sensor group includes a first pressure sensor 1 for monitoring the air pressure entering the stack, a second pressure sensor 2 for monitoring the air pressure leaving the stack, a third pressure sensor 3 for monitoring the hydrogen pressure entering the stack, a fourth pressure sensor 4 for monitoring the hydrogen pressure leaving the stack, a first temperature sensor 5 for monitoring the temperature of the coolant entering the stack, a second temperature sensor 6 for monitoring the temperature of the coolant leaving the stack, and a flow sensor 7 for monitoring the air flow entering the stack.

[0104] The monitoring unit can predict the water status of the fuel cell in real time, provide decision support for water management control, and adjust operating parameters in time to avoid flooding or membrane dry failure. Through accurate water status prediction and management, it can reduce fuel cell damage caused by improper water management and extend its service life.

[0105] Example 3

[0106] This embodiment provides a fuel cell management method, including: using the fuel cell internal water status monitoring method described in Example 1 to monitor the fuel motor to obtain the water content evaluation data in the fuel cell; the water content evaluation data in the fuel cell includes liquid water saturation and modal water content ;

[0107] When the modal water content of the anode in the fuel cell When the anode side membrane of the fuel cell is dry, it is determined that the anode side membrane of the fuel cell is dry. is the membrane dry threshold; controls the drainage frequency and exhaust frequency of the fuel cell minus the set value, increase the water content on the anode side, inhibit and restore the anode membrane dry failure.

[0108] When the modal water content of the cathode in the fuel cell When the cathode side membrane of the fuel cell is dry, the cathode air flow rate of the fuel cell is controlled to be less than the set value, weakening the convection effect of the fuel cell, increasing the water residence time on the cathode side, and inhibiting and recovering the cathode membrane dry failure.

[0109] When the liquid water saturation of the anode in the fuel cell When the anode side of the fuel cell is flooded, it is determined that the fuel cell has a flooding phenomenon; wherein, The membrane dry threshold is used to control the drainage frequency and exhaust frequency of the fuel cell to increase the set value, forming an intake pulse, promoting drainage, discharging excess water on the anode side, suppressing and recovering the anode flooding fault.

[0110] When the liquid water saturation of the cathode in the fuel cell When the fuel cell is flooded on the cathode side, the fuel cell is judged to have flooding; the cathode air intake flow of the fuel cell is controlled to increase by the set value, forming an intake pulse, enhancing convection, promoting drainage, discharging excess water on the anode side, suppressing and restoring the cathode flooding fault.

[0111] When the modal water content in the anode or cathode of the fuel cell The time is greater than the limit , the operating temperature of the fuel cell is controlled to be lowered to the set minimum value through the cooling device in the fuel cell; the condensation effect is strengthened to accelerate the recovery of membrane dry failure.

[0112] When the liquid water saturation in the anode or cathode of the fuel cell The time is greater than the limit , the operating temperature of the fuel cell is controlled to be lowered to the set maximum value through the cooling device in the fuel cell, thereby strengthening the evaporation effect and accelerating the recovery of water flooding failure.

[0113] Example 4

[0114] This embodiment provides an electronic device, including a storage medium and a processor; the storage medium is used to store instructions; the processor is used to operate according to the instructions to execute the fuel cell internal water status monitoring method described in Example 1.

[0115] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0116] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0117] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0118] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0119] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for monitoring the internal water status of a fuel cell, characterized in that: include: Monitor fuel cells to obtain real-time operating data; Input the real-time operation data into the pre-built fuel cell water state observation model to obtain the water content evaluation data in the fuel cell; The construction process of the fuel cell water state observation model includes: Obtaining the running storage data of the fuel cell, preprocessing the running storage data and adding real labels to obtain a training sample set; A fuel cell water state observation model is constructed based on a neural convolutional network, and the fuel cell water state observation model is trained using a training sample set to obtain training prediction data; the training loss value between the training prediction data and the true label is calculated based on component diffusion, liquid water transport, modal water transport, ion and electron transport, and heat transport in the direction perpendicular to the proton exchange membrane in the fuel cell; the Adam optimizer is used to optimize the parameters of the fuel cell water state observation model according to the training loss value, and the iterative training process is repeated until the set training cycle is reached to output the trained fuel cell water state observation model.

2. The method for monitoring the internal water status of a fuel cell according to claim 1, characterized in that: The real-time operation data and operation storage data of the fuel cell include stack voltage, stack current, stack inlet air pressure, stack outlet air pressure, stack inlet hydrogen pressure, stack outlet hydrogen pressure, stack inlet coolant temperature, stack outlet coolant temperature and stack inlet air flow.

3. The method for monitoring the internal water status of a fuel cell according to claim 1, characterized in that: The process of preprocessing running storage data includes: Check the missing values ​​in the running stored data and fill the missing values ​​in the running stored data through the interpolation algorithm; The outliers in the running storage data are identified by the Z-score algorithm, and the outliers are replaced by the mean or median in the running storage data; then the running storage data are normalized and standardized.

4. The method for monitoring the internal water status of a fuel cell according to claim 1, characterized in that: The training loss between the training prediction data and the true label is calculated based on the component diffusion, liquid water transport, modal water transport, ion and electron transport, and heat transport in the direction perpendicular to the proton exchange membrane in the fuel cell. The process includes: The fuel cell comprises anode and cathode bipolar plates, an anode flow channel, a cathode flow channel, an anode diffusion layer, a cathode diffusion layer, an anode microporous layer, a cathode microporous layer, a cathode catalyst layer, an anode catalyst layer and a proton exchange membrane; Calculate the data term loss between the training prediction data and the true label, expressed as: ； In the formula, is the data item loss function; Predict data items in the training data; is the data item in the real label; is the square sum function of absolute values; Calculate the boundary loss between the training prediction data and the true label, expressed as: ； In the formula, is the boundary loss, For training prediction data Gas composition at the interface; For training prediction data Liquid water components at the interface; For training prediction data Modal water component at the interface; The interface is the interface between the anode catalyst layer and the proton exchange membrane and the interface between the cathode catalyst layer and the proton exchange membrane; The interface is the interface between the anode catalyst layer and the anode microporous layer and the interface between the cathode catalyst layer and the cathode microporous layer; is the square sum function of absolute values; For training prediction data Liquid pressure at the boundary; For training prediction data Gas phase pressure at the boundary; For training prediction data Capillary pressure on the boundary; The boundaries are the interface between the anode flow channel and the anode diffusion layer and the interface between the cathode flow channel and the cathode diffusion layer; Calculate physical information losses based on component transport, liquid water transport, modal water transport, ion and electron transport, and heat transport in the direction perpendicular to the proton exchange membrane in the fuel cell; The training loss value is calculated based on the data item loss, boundary loss and physical information loss, and the expression formula is: ； In the formula, is the training loss value; is the physical information loss function; , , is the weight of each loss item.

5. The method for monitoring the internal water status of a fuel cell according to claim 4, characterized in that: The physical information loss calculated based on component diffusion, liquid water transport, modal water transport, ion and electron transport, and heat transport in the direction perpendicular to the proton exchange membrane in the fuel cell is expressed as: ； In the formula, is the diffusion state of the components; It is a liquid water transmission state; is the modal water transport state; It is the ion transport state; It is the state of electronic transmission; It is the state of heat transfer; The expression formula for calculating the diffusion state of the components is: ； In the formula, For time; are components, and the components are oxygen, hydrogen, nitrogen and water vapor; is the porosity; is the liquid water saturation; For each component The molar concentration of is the dimension perpendicular to the proton exchange membrane; For components The effective diffusion coefficient of For components The universal source term of ; The expression formula for calculating the liquid water transmission state is: ； In the formula, is the density of liquid water; is the intrinsic permeability of the porous medium; is the dynamic viscosity of liquid water; Source term generated for liquid water; The expression formula for calculating the modal water transmission state is: ； In the formula, is the dry film density; is the dry film equivalent; is the volume fraction of ionomer in the catalyst layer; is the modal water content; is the effective diffusion coefficient of modal water; Source terms generated for modal water; The expression formula for calculating the ion transport state is: ； In the formula, is the ionic conductivity of the proton exchange membrane; is the ionic potential; is the source term of the ionic potential; The expression formula for calculating the electron transfer state is: ； In the formula, is the intrinsic conductivity of the porous electrode skeleton; is the electron potential; is the source term of the electron potential; The expression formula for calculating the heat transfer state is: ； In the formula, is the equivalent heat capacity; is the equivalent thermal conductivity; is the temperature; is the general source term in the energy equation.

6. A fuel cell internal water status monitoring system, characterized in that: include: A monitoring unit, used to monitor the fuel cell and obtain real-time operating data; Input the real-time operation data into the pre-built fuel cell water state observation model to obtain the water content evaluation data in the fuel cell; An acquisition unit, used to acquire the operation storage data of the fuel cell, pre-process the operation storage data and add real labels to obtain a training sample set; Model building unit, building a fuel cell water state observation model based on a neural convolutional network; The training unit is used to train the fuel cell water state observation model using the training sample set to obtain training prediction data; calculate the training loss value between the training prediction data and the true label according to the component diffusion, liquid water transmission, modal water transmission, ion and electron transmission and heat transmission in the direction perpendicular to the proton exchange membrane in the fuel cell; use the Adam optimizer to optimize the parameters of the fuel cell water state observation model according to the training loss value, repeat the iterative training process until the set training cycle is reached to output the trained fuel cell water state observation model.

7. The fuel cell internal water status monitoring system according to claim 6, characterized in that: The monitoring unit is arranged in the fuel cell controller; the acquisition unit, the model building unit and the training unit are arranged in the cloud service platform; the fuel cell controller is configured with a data storage module, a terminal communication box and an OTA module; the fuel cell controller exchanges information through the terminal communication box; the OTA module is used to update the parameters of the fuel cell water state observation model; the data storage module is used to store the real-time operation data of the fuel cell and the water content evaluation data in the fuel cell.

8. The fuel cell internal water status monitoring system according to claim 6, characterized in that: The real-time operation data and operation storage data of the fuel cell include stack voltage, stack current, stack air pressure, stack air pressure, stack hydrogen pressure, stack hydrogen pressure, stack coolant temperature, stack coolant temperature and stack air flow; The monitoring unit is connected to a sensor group via a DC / DC converter; the sensor group includes a first pressure sensor for monitoring the air pressure entering the stack, a second pressure sensor for monitoring the air pressure leaving the stack, a third pressure sensor for monitoring the hydrogen pressure entering the stack, a fourth pressure sensor for monitoring the hydrogen pressure leaving the stack, a first temperature sensor for monitoring the coolant temperature entering the stack, a second temperature sensor for monitoring the coolant temperature leaving the stack, and a flow sensor for monitoring the air flow entering the stack.

9. A fuel cell management method, characterized in that: include: The fuel cell internal water status monitoring method according to any one of claims 1 to 5 is used to monitor the fuel motor to obtain the water content evaluation data in the fuel cell; the water content evaluation data in the fuel cell includes liquid water saturation and modal water content ; When the modal water content of the anode in the fuel cell When the anode side membrane of the fuel cell is dry, it is determined that the anode side membrane of the fuel cell is dry. is the membrane dry threshold; controls the drainage frequency and exhaust frequency of the fuel cell minus the set value; When the modal water content of the cathode in the fuel cell When the cathode side membrane of the fuel cell is dry, the cathode air flow rate of the fuel cell is controlled to be less than the set value; When the liquid water saturation of the anode in the fuel cell When the anode side of the fuel cell is flooded, it is determined that the fuel cell has a flooding phenomenon; wherein, The membrane dry threshold is used to control the drainage frequency and exhaust frequency of the fuel cell to increase the set value; When the liquid water saturation of the cathode in the fuel cell When the fuel cell is flooded on the cathode side, the fuel cell is judged to have flooding; the cathode air intake flow of the fuel cell is controlled to increase by the set value.

10. The fuel cell management method according to claim 9, characterized in that: When the modal water content in the anode or cathode of the fuel cell The time is greater than the limit , controlling the operating temperature of the fuel cell to drop to a set minimum value through a cooling device in the fuel cell; When the liquid water saturation in the anode or cathode of the fuel cell The time is greater than the limit , the operating temperature of the fuel cell is controlled to be lowered to the set maximum value through the cooling device in the fuel cell.

11. An electronic device comprising a storage medium and a processor; the storage medium is used to store instructions; characterized in that: The processor is used to operate according to the instructions to execute the method for monitoring the internal water status of a fuel cell according to any one of claims 1 to 5.

Images