CN115863712B

CN115863712B - Water management method and system for fuel cell

Info

Publication number: CN115863712B
Application number: CN202211463386.0A
Authority: CN
Inventors: 袁浩; 丁亚儒; 萧寒松
Original assignee: Suzhou Hydrogen New Energy Technology Co ltd
Current assignee: Suzhou Hydrogen New Energy Technology Co ltd
Priority date: 2022-11-22
Filing date: 2022-11-22
Publication date: 2023-11-28
Anticipated expiration: 2042-11-22
Also published as: CN115863712A; CN117558948A

Abstract

Embodiments of the present specification provide a water management method and system for a fuel cell, the method comprising: acquiring working data of a fuel cell; determining a normal condition or fault type of the fuel cell and a final solution based on the working data, wherein the final solution comprises a cathode pressure value solution and an anode pressure value solution of the fuel cell, the cathode pressure value solution is a pressure value corresponding to a cathode of the fuel cell when a preset condition is met, and the anode pressure value solution is a pressure value corresponding to an anode of the fuel cell when the preset condition is met; determining an adjustment mode based on the normal condition or the fault type; based on the adjustment mode, adjusting the cathode pressure value of the fuel cell to a cathode pressure value solution and adjusting the anode pressure value of the fuel cell to an anode pressure value solution.

Description

Water management method and system for fuel cell

Technical Field

The present disclosure relates to the field of fuel cells, and more particularly, to a method and system for water management of a fuel cell.

Background

At present, the fuel cell is used as a novel battery, has the advantages of high efficiency, low pollution, simple maintenance and the like, has an increasingly wide range in practical application, and needs to strictly control the internal parameters of the fuel cell to improve the performance and the utilization rate of the fuel cell due to the particularity of the fuel cell. And the internal water content of the fuel cell is one of the main factors affecting the performance and service life of the fuel cell itself.

It is therefore desirable to provide a method and system for water management of a fuel cell that allows for real-time monitoring and management of the internal water content of the fuel cell, which helps to improve the performance and utilization of the cell.

Disclosure of Invention

One of the embodiments of the present specification provides a water management method for a fuel cell. The water management method of the fuel cell comprises the following steps: acquiring working data of a fuel cell; determining a normal condition or fault type and a final solution of the fuel cell based on the working data, wherein the final solution comprises a cathode pressure value solution and an anode pressure value solution of the fuel cell, the cathode pressure value solution is a pressure value corresponding to a cathode of the fuel cell when a preset condition is met, and the anode pressure value solution is a pressure value corresponding to an anode of the fuel cell when the preset condition is met; determining an adjustment mode based on the normal condition or the fault type; and adjusting a cathode pressure value of the fuel cell to the cathode pressure value solution and adjusting an anode pressure value of the fuel cell to the anode pressure value solution based on the adjustment mode.

One of the embodiments of the present specification provides a water management system of a fuel cell, the water management system of the fuel cell including: an acquisition module for acquiring operation data of the fuel cell; a first determining module, configured to determine, based on the operation data, a normal condition or a fault type of the fuel cell and a final solution, where the final solution includes a cathode pressure value solution and an anode pressure value solution of the fuel cell, where the cathode pressure value solution is a pressure value corresponding to a cathode of the fuel cell when a preset condition is met, and the anode pressure value solution is a pressure value corresponding to an anode of the fuel cell when the preset condition is met; a second determining module for determining an adjustment mode based on the normal condition or the fault type; and the adjusting module is used for adjusting the cathode pressure value of the fuel cell to the cathode pressure value solution and adjusting the anode pressure value of the fuel cell to the anode pressure value solution based on the adjusting mode.

One of the embodiments of the present specification provides a water management device for a fuel cell, the device comprising at least one processor and at least one memory; the at least one memory is configured to store computer instructions; the at least one processor executes at least some of the computer instructions to implement a method of water management for a fuel cell.

One of the embodiments of the present specification provides a computer-readable storage medium storing computer instructions that, when read by a computer in the storage medium, the computer performs a method of water management of a fuel cell.

Drawings

The present specification will be further elucidated by way of example embodiments, which will be described in detail by means of the accompanying drawings. The embodiments are not limiting, in which like numerals represent like structures, wherein:

FIG. 1 is a schematic illustration of an application scenario of a water management system for a fuel cell according to some embodiments of the present disclosure;

FIG. 2 is an exemplary block diagram of a water management system for a fuel cell according to some embodiments of the present disclosure;

FIG. 3 is an exemplary flow chart of a method of water management for a fuel cell according to some embodiments of the present description;

FIG. 4 is an exemplary schematic diagram of a fault model shown in accordance with some embodiments of the present description;

FIG. 5 is an exemplary flow chart for determining a final solution based on work data according to some embodiments of the present description;

FIG. 6 is an exemplary schematic diagram of a power prediction model shown in accordance with some embodiments of the present description;

FIG. 7 is an exemplary schematic diagram of candidate combined iterative updates shown in accordance with some embodiments of the present description.

Detailed Description

In order to more clearly illustrate the technical solutions of the embodiments of the present specification, the drawings that are required to be used in the description of the embodiments will be briefly described below. It is apparent that the drawings in the following description are only some examples or embodiments of the present specification, and it is possible for those of ordinary skill in the art to apply the present specification to other similar situations according to the drawings without inventive effort. Unless otherwise apparent from the context of the language or otherwise specified, like reference numerals in the figures refer to like structures or operations.

It will be appreciated that "system," "apparatus," "unit" and/or "module" as used herein is one method for distinguishing between different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the words can be replaced by other expressions.

As used in this specification and the claims, the terms "a," "an," "the," and/or "the" are not specific to a singular, but may include a plurality, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.

A flowchart is used in this specification to describe the operations performed by the system according to embodiments of the present specification. It should be appreciated that the preceding or following operations are not necessarily performed in order precisely. Rather, the steps may be processed in reverse order or simultaneously. Also, other operations may be added to or removed from these processes.

Fig. 1 is a schematic view of an application scenario of a water management system of a fuel cell according to some embodiments of the present disclosure.

As shown in fig. 1, the application scenario 100 of the water management system of the fuel cell may include a fuel cell 110, a memory 120, a processor 130, a user terminal 140, and a network 150.

In some embodiments, one or more components in the application scenario 100 of the water management system of the fuel cell may communicate data to other components over the network 150. For example, the processor 130 may obtain information and/or data in the user terminal 140, the fuel cell 110, and the memory 120 through the network 150, or may send information and/or data to the user terminal 140 and the memory 120 through the network 150.

The fuel cell 110 may include various types of cells. For example, the fuel cell 110 may be a solid oxide fuel cell, a phosphoric acid fuel cell, a proton exchange membrane fuel cell, or the like. The fuel cell 110 may include a fuel cell stack, a hydrogen supply system, an air system, a water thermal management system, an electronic control system, and the like. The water thermal management system may include a water management system and a thermal management system. Among other things, water management can be used to provide reasonable water content in the membrane electrode to ensure good conduction of hydrogen ions in the membrane. Thermal management can control the operating temperature of the fuel cell to a safe and reasonable range. In the fuel cell 110, sensors such as a pressure sensor, a temperature sensor, and the like are used for the hydrogen gas supply system, the air system, and the water heat management system. The processor 130 may obtain fuel cell operation data via sensors in the subsystem.

The memory 120 may be used to store data and/or instructions related to the application scenario 100 of the water management system of the fuel cell. For example, the memory 120 may store the fuel cell's own data (e.g., electrolyte type, electrolyte concentration, proton exchange membrane material, etc.). The own data of the fuel cell may be stored in the memory 120 in advance. In some embodiments, the memory 120 may store data and/or information obtained from the fuel cell 110, the processor 130, etc. For example, the memory 120 may store voltage data, current data, temperature data, cathode pressure values, anode pressure values, and the like of the fuel cell.

Memory 120 may include one or more memory components, each of which may be a separate device or may be part of another device. In some embodiments, memory 120 may include Random Access Memory (RAM), read Only Memory (ROM), mass storage, removable memory, volatile read-write memory, and the like, or any combination thereof. By way of example, mass storage may include magnetic disks, optical disks, solid state disks, and the like. In some embodiments, the memory 120 may be implemented on a cloud platform.

Processor 130 may process data and/or information obtained from other devices or system components. The processor may execute program instructions to perform one or more of the functions described in this disclosure based on such data, information, and/or processing results. For example, the processor 130 may be configured to determine the normal condition or fault type and the final solution of the fuel cell based on the operational data. For another example, the processor 130 may be configured to determine an adjustment mode based on a normal condition or a type of fault, adjust a cathode pressure value of the fuel cell to a cathode pressure value solution based on the adjustment mode, and adjust an anode pressure value of the fuel cell to an anode pressure value solution.

In some embodiments, processor 130 may contain one or more sub-processing devices (e.g., single-core processing devices or multi-core processing devices). By way of example only, the processor 130 may include a Central Processing Unit (CPU), an Application Specific Integrated Circuit (ASIC), an Application Specific Instruction Processor (ASIP), a Graphics Processor (GPU), a Physical Processor (PPU), a Digital Signal Processor (DSP), a Field Programmable Gate Array (FPGA), an editable logic circuit (PLD), a controller, a microcontroller unit, a Reduced Instruction Set Computer (RISC), a microprocessor, or the like, or any combination thereof.

User terminal 140 refers to one or more terminal devices or software used by a user. The user terminal 140 may include a processing unit, a display unit, an input/output unit, a sensing unit, a storage unit, and the like. The sensing unit may include, but is not limited to, a light sensor, a distance sensor, an acceleration sensor, a gyro sensor, a sound detector, etc., or any combination thereof. In some embodiments, user terminal 140 may be one or any combination of a mobile device, tablet computer, laptop computer, or other input and/or output enabled device. In some embodiments, one or more users of the user terminal 140 may be used, including users who directly use the service, as well as other related users. In some embodiments, the mobile device may be a smart phone, smart watch, or the like. In some embodiments, the user may refer to a fuel cell servicer or other service user. The above examples are only intended to illustrate the broad scope of the user terminal 140 devices and not to limit the scope thereof.

Network 150 may connect components of the system and/or connect the system with external resource components. Network 150 enables communication between the various components and with other components outside the system to facilitate the exchange of data and/or information. For example, processor 130 may retrieve fuel cell operating data from memory 120 via network 150.

In some embodiments, network 150 may be any one or more of a wired network or a wireless network. For example, the network 150 may include a cable network, a fiber optic network, a telecommunications network, the internet, a Local Area Network (LAN), a Wide Area Network (WAN), a Wireless Local Area Network (WLAN), a Metropolitan Area Network (MAN), a Public Switched Telephone Network (PSTN), a bluetooth network, a ZigBee network, a Near Field Communication (NFC), an intra-device bus, an intra-device line, a cable connection, and the like, or any combination thereof. The network connection between the parts can be in one of the above-mentioned ways or in a plurality of ways. In some embodiments, the network may be a point-to-point, shared, centralized, etc. variety of topologies or a combination of topologies.

It should be noted that the application scenario 100 of the water management system of a fuel cell is provided for illustrative purposes only and is not intended to limit the scope of the present application. Many modifications and variations will be apparent to those of ordinary skill in the art in light of the present description. For example, the application scenario 100 of the water management system of a fuel cell may implement similar or different functions on other devices. However, such changes and modifications do not depart from the scope of the present application.

Fig. 2 is an exemplary block diagram of a water management system for a fuel cell according to some embodiments of the present description.

In some embodiments, the water management system 200 of the fuel cell may include an acquisition module 210, a first determination module 220, a second determination module 230, and a conditioning module 240. The water management system 200 of the fuel cell according to the embodiment of the present specification will be described in detail below. It should be noted that the following examples are only for explaining the present specification, and do not constitute a limitation of the present specification.

As shown in fig. 2, the water management system 200 of the fuel cell may include:

in some embodiments, the acquisition module 210 may be used to acquire operating data for the fuel cell. For more description of acquiring operational data, see fig. 3 and its associated description.

In some embodiments, the first determining module 220 may be configured to determine, based on the operation data, a normal condition or a fault type of the fuel cell and a final solution, where the final solution includes a cathode pressure value solution and an anode pressure value solution of the fuel cell, the cathode pressure value solution being a pressure value corresponding to a cathode of the fuel cell when a preset condition is satisfied, and the anode pressure value solution being a pressure value corresponding to an anode of the fuel cell when the preset condition is satisfied. For more explanation on determining the normal condition or failure type and final solution of the fuel cell, see fig. 3 and its associated description.

In some embodiments, the first determination module 220 may be further configured to determine a battery time period characteristic based on the operational data; based on the battery time period characteristics, a normal condition or fault type is determined. For more explanation about the normal condition or type of failure, see fig. 3 and its associated description.

In some embodiments, the first determination module 220 may be further configured to determine the fault type through a fault model based on the battery time period characteristics, wherein the fault model is a machine learning model. For more explanation of the fault model, see fig. 4 and its associated description.

In some embodiments, the fault model may further include a first embedded layer through which the battery time period characteristics are obtained.

In some embodiments, the first determination module 220 may be further configured to determine a plurality of candidate solutions based on the working data; determining a battery output power predicted value corresponding to each candidate solution in the plurality of candidate solutions through a power prediction model based on the plurality of candidate solutions and the working data, wherein the power prediction model is a machine learning model; and determining a final solution based on the battery output power predicted value, wherein the final solution is a corresponding candidate solution of which the battery output power predicted value meets a preset condition. For more description of the power prediction model, see fig. 6 and its associated description.

In some embodiments, the power prediction model may include a second embedded layer and a power prediction layer, the input to the power prediction layer further including a battery period feature.

In some embodiments, the input to the power prediction layer further includes a fault type. In some embodiments, the second embedded layer shares parameters with the first embedded layer of the fault model.

In some embodiments, the first determining module 220 may be further configured to determine the final solution through a preset algorithm based on the working data. For more explanation of the preset algorithm, see fig. 7 and its associated description.

In some embodiments, the preset algorithm may include: generating a plurality of candidate combinations, wherein the candidate combinations include cathode pressure values and anode pressure values of the fuel cells; performing multiple rounds of iterative updating on the multiple candidate combinations to determine a final solution; at least one of the multiple iterations includes: and updating the candidate combination based on the pressure value change amplitude corresponding to the relation update of the historical optimal combination for at least one candidate combination, wherein the historical optimal combination is determined based on an evaluation algorithm.

In some embodiments, the plurality of candidate combinations are candidate solutions for which the battery output power prediction value satisfies a set condition, wherein the set condition includes a set threshold value that is less than a preset threshold value, wherein the preset threshold value is a threshold value in the preset condition.

In some embodiments, the historical optimal combination includes an independent optimal combination corresponding to the candidate combination, and an associated optimal combination that collectively corresponds to the plurality of candidate combinations.

In some embodiments, the evaluation algorithm may further comprise: determining a battery output power predicted value corresponding to each candidate combination in the plurality of candidate combinations through a power prediction model based on the plurality of candidate combinations and the working data; based on the battery output power predictions, a historical optimal combination is determined.

A second determination module 230 for determining an adjustment mode based on a normal condition or a fault type. For more description of determining the adjustment mode, see fig. 3 and its associated description.

The adjusting module 240 is configured to adjust a cathode pressure value of the fuel cell to a cathode pressure value solution and adjust an anode pressure value of the fuel cell to an anode pressure value solution based on the adjustment mode. For more explanation of adjusting the pressure values to the pressure value solutions, see fig. 3 and its associated description.

It should be understood that the system shown in fig. 2 and its modules may be implemented in a variety of ways.

It should be noted that the above description of the water management system of the fuel cell and its modules is for convenience of description only and is not intended to limit the present description to the scope of the illustrated embodiments. It will be appreciated by those skilled in the art that, given the principles of the system, various modules may be combined arbitrarily or a subsystem may be constructed in connection with other modules without departing from such principles. In some embodiments, the acquisition module 210, the first determination module 220, the second determination module 230, and the adjustment module 240 disclosed in fig. 2 may be different modules in one system, or may be one module to implement the functions of two or more modules described above. For example, each module may share one memory module, or each module may have a respective memory module. Such variations are within the scope of the present description.

Fig. 3 is an exemplary flow chart of a method of water management for a fuel cell according to some embodiments of the present description. In some embodiments, the process 300 may be performed by the processor 130. As shown in fig. 3, the process 300 may include the steps of:

in step 310, operation data of the fuel cell is acquired. In some embodiments, step 310 may be performed by the acquisition module 210.

The operation data may refer to data about the fuel cell in an operating state. The operating data may include data related to the voltage, current, temperature, cathode pressure, anode pressure, self-data of the fuel cell (e.g., electrolyte type, electrolyte concentration, proton exchange membrane material, etc.), etc. For example, the electrolyte species of the fuel cell may be a perfluorosulfonic acid type solid polymer. The operational data may include data relating to the fuel cell at a plurality of different points in time. The plurality of different time points may be the current time point and a plurality of different time points of a period of time prior to the current time point. Each of the plurality of different time points has corresponding data relating to the fuel cell.

In some embodiments, the acquisition module 210 may acquire the operating data of the fuel cell in a variety of ways. For example, the acquisition module 210 may acquire operational data of the fuel cell via sensor acquisition. Illustratively, a voltage sensor collects voltage data, a current sensor collects current data, a temperature sensor collects temperature data, a pressure sensor collects cathode pressure values, anode pressure values, and the like. For another example, the acquisition module 210 may acquire the own data of the fuel cell through the memory 120.

Step 320, based on the operational data, determines the fuel cell's normal condition or fault type and final solution. In some embodiments, step 320 may be performed by the first determination module 220.

The normal condition may refer to a condition in which the operation data of the fuel cell is normal. In some embodiments, the normal operation data of the fuel cell may refer to the operation data of the voltage, current, temperature, cathode pressure value, anode pressure value, etc. of the cell are all within the normal range. The normal range may be preset in advance.

The fault type may guide the kind of fault cause that the battery cannot normally operate. For example, fault types may include flooding, film drying, and other types. Flooding may refer to the excessive internal water content of the cell. Film drying may refer to a too low water content within the cell. Other types may include a failure type such as starvation of the battery due to a long period of non-charging causing the battery to be non-rechargeable.

In some embodiments, the first determination module 220 may determine the fault type based on the operational data. For example, the normal condition or the type of failure is determined according to whether the operation data is within the normal range. For another example, the first determination module 220 may determine a normal condition or a failure type, etc., through a machine learning model based on the operational data.

In some embodiments, the first determination module 220 may determine a battery period characteristic based on the operational data and determine a normal condition or a type of fault of the fuel cell based on the battery period characteristic.

The battery time period characteristic may represent characteristics of operation data of the fuel cell at a plurality of different time points of a period of time preceding the current time point. In some embodiments, the battery time period characteristics may include a voltage maximum (e.g., voltage maximum, voltage minimum), a current maximum (e.g., current maximum, current minimum), a temperature maximum (e.g., temperature maximum, temperature minimum), a maximum voltage rate of change, a maximum temperature rate of change, etc., for a period of time prior to the current point in time. In some embodiments, the maximum voltage change rate refers to the maximum value of the ratio of the amount of change in voltage to the time it takes to complete the change. The maximum temperature change rate means the maximum value of the ratio of the amount of temperature change to the time taken to complete the change.

In some embodiments, the battery time period characteristics may be determined based on partial operational data of a period of time prior to the current point in time. For example, the voltage of the fuel cell is measured in real time by a voltage sensor, and the voltage maximum value and the voltage minimum value in the battery period characteristic of the previous period of time at the present point of time can be obtained.

In some embodiments, the first determination module 220 may obtain the battery period characteristics based on a first embedded layer of the fault model. See in particular fig. 4 for an illustration of the failure model and the first embedded layer.

In some embodiments, battery health or fault type may be determined based on processing of the time-domain features by a machine learning model (e.g., a fault model, etc.). The description of the fault model is specifically referred to in fig. 4.

In some embodiments of the present disclosure, the working data is used to determine the time period characteristics of the fuel cell, and the normal condition or the fault type of the fuel cell is determined according to the time period characteristics of the fuel cell, so that the accuracy of the determined normal condition or fault type can be improved, which is beneficial to the subsequent adjustment of the fuel cell, and the performance and the utilization rate of the fuel cell are improved.

The final solution may refer to a set of pressure values that bring the battery to an optimal performance state under the current conditions. In some embodiments, the final solution may include a cathode pressure value solution and an anode pressure value solution of the fuel cell. In some embodiments, the cathode pressure value is solved as the pressure value corresponding to the cathode of the fuel cell when the preset condition is satisfied. The anode pressure value is solved into a pressure value corresponding to the anode of the fuel cell when the preset condition is satisfied. The preset condition may refer to a condition to be satisfied preset in advance. In some embodiments, the preset conditions may be set according to circumstances. In some embodiments, the preset condition may include the battery output power prediction value being greater than and/or equal to a preset threshold. The preset threshold may refer to a minimum value of the battery output power prediction value. The preset threshold may be set manually. The battery output power prediction value may be determined by a power prediction model. For more description of battery output power predictions and power prediction models, see the associated description of fig. 5 and 6.

In some embodiments, the first determination module 220 may determine the final solution based on the working data. For example, the first determination module 220 may determine a set of cathode pressure values, anode pressure values for which the cell reaches an optimal performance state based on the current point in time and data related to the fuel cell at a plurality of different points in time that are a period of time prior to the current point in time. The first determination module 220 may determine a set of cathode pressure values, anode pressure values for which the cell reaches an optimal performance state as a final solution. The optimal performance state may refer to the battery output power prediction value being the largest. The battery output power prediction value may be determined by a power prediction model. For more description of battery output power predictions and power prediction models, see the associated description of fig. 5 and 6.

In some embodiments, the optimal performance state varies continuously with the condition of the fuel cell. The first determination module 220 may determine the final solution in real-time.

In some embodiments, the first determination module 220 may determine the final solution through a preset algorithm based on the working data.

In some embodiments, the first determination module 220 may determine the final solution by a variety of preset algorithms based on the working data. The preset algorithm is an algorithm preset to determine the final solution. For example, the preset algorithm may include a regression analysis method, a discriminant analysis method, a random search method, and the like.

In some embodiments, the first determination module 220 may determine the final solution by iteratively updating a plurality of candidate combinations, with particular reference to fig. 7 for an illustration of the iterative updating of candidate combinations to determine the final solution.

In some embodiments of the present disclosure, the direction of exploration in the solution space may be dynamically adjusted by a preset algorithm in combination with the working data, so as to more quickly determine the final solution.

In some embodiments, the first determination module 220 may determine a plurality of candidate solutions based on the operational data and determine a battery output power prediction value via a power prediction model based on the plurality of candidate solutions and the operational data, thereby determining a final solution. The description of the above is specifically referred to fig. 5.

Step 330, based on the normal condition or the type of fault, a regulation mode is determined. In some embodiments, step 330 may be performed by the second determination module 230.

The regulation mode can be used for indicating different running states of the battery and carrying out different modes of regulating the water content in the battery. The regulation mode may include a normal regulation mode, a flooding regulation mode, and a membrane dry regulation mode. The normal regulation mode may refer to a regulation mode when the battery is operating normally. The flooding adjustment mode may refer to an adjustment mode when the fuel cell enters a flooding failure. The membrane dry conditioning mode may refer to a conditioning mode when the fuel cell enters a membrane dry failure.

In some embodiments, when the internal water content of the battery exceeds a maximum water content threshold or the water saturation of the gas diffusion layer exceeds a maximum saturation threshold, the battery is judged to have a flooding fault, and the regulation mode is determined to be a flooding regulation mode.

In some embodiments, when the internal water content of the battery is lower than a water content minimum threshold or the water saturation of the gas diffusion layer is lower than a saturation minimum threshold, the battery is judged to have a membrane dry fault, and the regulation mode is determined to be a membrane dry regulation mode.

In some embodiments, the normal adjustment manner may take the temperature data in the working data at the last time point in the current sampling period as the temperature at the current time point, determine, through a power prediction model, the predicted values of the battery output power corresponding to the working data at different time points in the current sampling period, and determine the cathode pressure value and the anode pressure value corresponding to the maximum predicted values of the battery output power as the final solution. The current sampling period may be set manually. For more description of the power prediction model, see related description of other parts. In some embodiments, the cathode pressure value and the anode pressure value of the fuel cell may be adjusted for a normal fuel cell at intervals, and the specific interval may be manually set.

Step 340, based on the adjustment mode, adjusting the cathode pressure value of the fuel cell to a cathode pressure value solution and adjusting the anode pressure value of the fuel cell to an anode pressure value solution. In some embodiments, step 340 may be performed by adjustment module 240.

The cathode pressure value may refer to the pressure value of the cell cathode at the current point in time. In some embodiments, adjustments are made for different fault types, a cathode pressure value solution for the current cell condition is found, and the cathode pressure value of the final solution is configured at the cathode. In some embodiments, the cathode pressure value may be adjusted by the air compressor and throttle.

The anode pressure value may refer to the pressure value of the battery anode at the current point in time. In some embodiments, adjustments are made for different fault types, an anode pressure value solution for the current cell condition is found, and the anode pressure value of the final solution is configured at the anode. In some embodiments, the anode pressure value may be determined by both the flow valve and the hydrogen circulation pump.

In some embodiments, for flooding faults, when the internal water content is high, two-pole air supply is stopped, the internal water content is slowly consumed, the output power is slowly increased, when the output power starts to decrease, the temperature at the inflection point is selected as the temperature of the current time point, and the pressure value of the fuel cell is adjusted to be the final solution based on the final solution determined by a power prediction model and the like.

In some embodiments, for a dry membrane failure, where the internal water content is low, a high pressure value is manually selected to be configured at two poles, the internal water content is slowly increased, the output power is slowly increased, when the output power begins to decrease, the temperature at the inflection point is selected as the temperature at the current time point, and the pressure value of the fuel cell is adjusted to the final solution based on a final solution determined by a power prediction model or the like.

In some embodiments, for normal conditions, the conditioning module 240 may configure the final solutions determined under normal conditions to the cathode and anode of the fuel cell, respectively.

In some embodiments of the present disclosure, the internal water content of the fuel cell is monitored in real time, so that the fuel cell is timely adjusted for normal conditions or different fault types, which is helpful to improve the performance and the utilization rate of the fuel cell.

FIG. 4 is an exemplary schematic diagram of a fault model shown in accordance with some embodiments of the present description.

As shown in fig. 4, the fault model 400 may include a first embedded layer 420, a fault output layer 440, and the like.

The failure model may refer to a machine learning model that determines battery failure. In some embodiments, the fault model may be a trained machine learning model. For example, the fault model may include any one or combination of a recurrent neural network model, a convolutional neural network, or other custom model structure, etc.

The first embedded layer 420 may be used to extract battery period characteristics. In some embodiments, the model type of the first embedded layer 420 may include, but is not limited to, a deep neural network model. In some embodiments, the input of the first embedded layer 420 may be the operational data 410 and the output may be the battery time period feature 430.

The fault output layer 440 may be used to determine a normal condition or fault type. In some embodiments, the model type of the fault output layer 440 may include, but is not limited to, a deep neural network model. In some embodiments, the input of the fault output layer 440 may be the battery period feature 430 and the output may be the normal condition or fault type 450.

In some embodiments, the first embedded layer may be determined based on the embedded layer sample and the embedded layer tag. In some embodiments, the embedded layer sample may include historical operational data and the embedded layer tag may include historical battery period characteristics. The training process comprises the following steps: and inputting an embedded layer sample with an embedded layer label into a first embedded layer, and updating parameters of the first embedded layer through training until the conditions that a loss function is smaller than a threshold value, convergence is achieved or a training period reaches the threshold value are met.

In some embodiments, the fault model may be determined by first embedded layer, fault output layer joint training. In some embodiments, each of the first training samples of the fault model may include historical operating data of the fuel cell when the history fails. In some embodiments, each of the first training tags of the failure model may include film dry, flooded, and others. The labels may be labeled based on manual labeling or other viable means. In some embodiments, the first training sample and the first training tag may be obtained based on related data generated during historical operation of the battery. The embedded layer samples and the embedded layer labels may be in a one-to-one correspondence with the first training samples and the first training labels. For example, each of the plurality of training samples and labels is the same historical working data. The battery time period characteristics of the first embedded layer output may be input to the fault output layer. The process of joint training may include: taking the historical working data in the embedded layer sample as the input of the first embedded layer; taking the battery time period characteristics output by the first embedded layer as the input of a fault output layer to determine the output of the fault output layer; constructing a loss function by the normal condition or the fault type output by the fault output layer and the first training label; and iteratively updating based on the loss function until the condition that the loss function is smaller than a threshold value, converged or the training period reaches the threshold value is met, and obtaining a trained first embedded layer and a fault output layer.

In some embodiments, the first embedding layer and fault output layer joint training may promote accuracy of the obtained fault model output.

In some embodiments of the present description, the fault type is determined by a fault model, and the adjustment mode is determined according to the fault type, so that accuracy of the determined adjustment mode is guaranteed. The fault type determined by the fault model can also be used as the input of the power prediction model, which is further beneficial to ensuring the accuracy of the determined final solution.

FIG. 5 is an exemplary flow chart for determining a final solution based on work data according to some embodiments of the present description. In some embodiments, the process 500 may be performed by the processor 130. The process 500 may include the steps of:

step 510, determining a plurality of candidate solutions based on the working data.

The candidate solution may refer to a set of candidate cathode pressure value solutions and candidate anode pressure value solutions to be selected as the final solution. The candidate cathode pressure value solution may refer to a cathode pressure value solution to be selected as the final solution. The candidate anode pressure value solution may refer to an anode pressure value solution to be selected as the final solution. In some embodiments, the number of candidate solutions may be multiple, each containing a corresponding set of candidate cathode pressure value solutions and candidate anode pressure value solutions. For example, the candidate solution 1 and the candidate solution 2 may include the candidate solution 1 and the candidate solution 2, where the candidate solution for the cathode pressure value and the candidate solution for the anode pressure value in the candidate solution 1 are 0.3MPa and 0.1MPa, respectively.

In some embodiments, the processor 130 may preset a candidate solution database, and determine a candidate solution based on the operating data of the fuel cell at the current time and the candidate solution database. The operation data of the fuel cell at the current time may be represented by the current cell period feature vector. The current cell period feature vector may represent a period feature of the fuel cell at the current time. Each element in the current cell period feature vector may represent a voltage value, a current value, a temperature value, a voltage change rate, a temperature change rate, and the like of the fuel cell at the current time. For example, the current cell period feature vector is (20, 5, 70,5%, … …), which indicates that the voltage value of the fuel cell at the current time is 20V, the current value is 5A, the temperature value is 70 ℃, the voltage change rate is 5%, the temperature change rate is 5%, and the like. In some embodiments, the candidate solution database may include multiple sets of historical operating data for the fuel cell. In some embodiments, each set of historical operating data in the candidate solution database includes a historical voltage, a historical current, a historical temperature, a historical cathode pressure value, a historical anode pressure value, and the like, corresponding to different historical points in time. Each set of historical working data in the candidate solution database may be represented by a historical battery period feature vector.

In some embodiments, the processor 130 may determine the candidate solution by comparing the similarity between the current battery period feature vector and the historical battery period feature vectors in the candidate solution database. In some embodiments, the similarity between the current battery period feature vector and the historical battery period feature vector may be represented by a vector distance. Vector distances may include euclidean distances, geometric distances, and the like. The smaller the vector distance, the higher the similarity between the current battery period feature vector and the historical battery period feature vector. In some embodiments, the processor 130 may select a plurality of historical battery period feature vectors having a similarity to the current battery period feature vector that is above a similarity threshold. The processor 130 may consider the historical cathode pressure value and the historical anode pressure value corresponding to each historical battery period feature vector above the similarity threshold as a candidate cathode pressure value solution and a candidate anode pressure value solution, respectively, in one candidate solution. A plurality of candidate solutions are determined from a plurality of historical battery period feature vectors above a similarity threshold. The similarity threshold may be set manually. For example, the candidate solution database includes the historical battery period feature vector A, B, C, etc., the similarity threshold is 90%, and the similarity between the historical battery period feature vector A, B, C, etc. and the current battery period feature vector a is 80%, 91% and 98%, etc., respectively, and then the candidate solution includes the historical battery period feature vector B, C, etc.

In some embodiments, when the number of historical battery period feature vectors in the candidate solution database that are similar to the current battery period feature vector is small, the processor 130 may randomly generate a plurality of sets of candidate solutions, each set of candidate solutions including a respective candidate cathode pressure value solution and candidate anode pressure value solution.

Step 520, determining, by the power prediction model, a battery output power prediction value corresponding to each of the plurality of candidate solutions based on the plurality of candidate solutions and the working data.

The battery output power prediction value may be a power value obtained by predicting the battery output power corresponding to each candidate solution. For example, the battery output power prediction value corresponding to the candidate solution M obtained by the power prediction model is 100W.

The power prediction model may be used to predict battery output power predictions for the candidate solutions. In some embodiments, the power prediction model may be a machine learning model, e.g., a deep neural network model, a convolutional network model, or the like.

In some embodiments, the input to the power prediction model may include candidate solutions and working data. The operation data may include the temperature of the fuel cell, the data of the fuel cell itself, and the like. For more explanation of the working data, see the relevant description of fig. 3. The output of the power prediction model may include a battery output power prediction value for each candidate solution. The processor 130 may determine a battery output power prediction value corresponding to each of the plurality of candidate solutions by inputting each of the plurality of candidate solutions and the working data into the power prediction model, respectively.

In some embodiments, the power prediction model may be obtained through training. In some embodiments, the power prediction model may be obtained by training an initial power prediction model. And inputting a second training sample into the initial power prediction model to obtain a predicted value of the output power of the historical battery. In the training process, a loss function is established based on the output results of the second label and the initial power prediction model, and parameters of the initial power prediction model are iteratively updated based on the loss function until training conditions are met and training is completed, so that a trained power prediction model is obtained. Training conditions may include loss function convergence, etc.

In some embodiments, the corresponding second training samples may include multiple sets of historical work data and corresponding historical candidate solutions, and the second labels of the second training samples may be corresponding historical battery output power actual values. The second label may be manually labeled.

In some embodiments, the power prediction model may include a second embedded layer and a power prediction layer. And determining a battery output power predicted value corresponding to each candidate solution through the second embedding layer and the power predicting layer. Details regarding the second embedded layer and the power prediction layer may be found in the description of the other contents of the present specification, for example, fig. 6.

In step 530, a final solution is determined based on the predicted battery output power value, where the final solution is a corresponding candidate solution for which the predicted battery output power value satisfies a preset condition.

In some embodiments, the processor 130 may determine a corresponding candidate solution for which the battery output power prediction value satisfies the preset condition as the final solution. For example, if the predicted values of the battery output power corresponding to the candidate solution X, Y, Z are 90W, 100W, 150W, etc., respectively, and the preset threshold is preset to 120W, the final solution is the candidate solution Z. In some embodiments, the preset condition may include a maximum battery output power prediction value. The processor 130 may determine a candidate solution, of the plurality of candidate solutions, for which the battery output power prediction value is the largest, as the final solution. For more explanation of the preset conditions and the preset threshold values, see the relevant description of fig. 3.

In some embodiments of the present disclosure, a plurality of candidate solutions are determined based on the operational data, and then a more accurate predicted value of battery output power corresponding to each of the plurality of candidate solutions may be predicted by a power prediction model. The final solution is determined based on the output power predicted value, so that the determined final solution can be accurate, and the output power of the fuel cell is high. The working state of the fuel cell is continuously changed, and the fuel cell can always keep higher output power by continuously adjusting the anode-cathode voltage value solution and the anode-cathode voltage value solution, so that the performance and the utilization rate of the fuel cell are improved.

FIG. 6 is an exemplary schematic diagram of a power prediction model shown in accordance with some embodiments of the present description.

As described in fig. 6, the power prediction model may include a second embedded layer 610 and a power prediction layer 640.

In some embodiments, the second embedded layer 610 may be used to obtain the battery period feature 430. The second embedded layer 610 may be a machine learning model, for example, the second embedded layer 610 may include convolutional neural networks, and the like. In some embodiments, the input of the second embedded layer 610 may include the operational data 410 and the output may include the battery period feature 430. See the relevant description of fig. 3 for working data. See the relevant description of fig. 3 and 4 for battery time period features. Further, the processor 130 may determine the final solution 660 based on the battery output power prediction value 650. For details on the final solution, see the description of fig. 5 of the present specification.

In some embodiments of the present description, the redundant data or the remaining irrelevant data in the working data may be removed by acquiring the battery period feature through the second embedded layer 610. For example, the occupied characters in the working data can be removed, the storage space of the data can be saved, and the operation efficiency of the data can be improved.

In some embodiments, the power prediction layer 640 may be configured to obtain a battery output power prediction value 650 corresponding to each candidate solution. The power prediction layer 640 may be implemented by, for example, convolutional Neural networks (Convolutional Neural Network, CNN), neural Networks (NN), or the like. In some embodiments, the inputs to the power prediction layer 640 may include the working data 410, the candidate solution 620, and the like. The output of the power prediction layer 640 may include a battery output power prediction value 650 for each candidate solution.

In some embodiments, the power prediction model may be obtained through training. In some embodiments, the power prediction model may be obtained by jointly training the initial second embedded layer and the initial power prediction layer. Inputting a third training sample into the initial second embedded layer to obtain historical battery time period characteristics; the historical battery time period characteristics, the plurality of historical candidate solutions, and the historical operational data are then used as inputs to an initial power prediction layer. In the training process, a loss function is established based on the output results of the third tag and the initial power prediction layer, and parameters of the initial second embedded layer and the initial power prediction layer are iteratively updated based on the loss function at the same time until the preset condition is met, and training is completed. Parameters of the second embedded layer and the power prediction layer after training is completed can also be determined, and a trained power prediction model is obtained.

In some embodiments, the third training sample may include historical work data and a plurality of historical candidate solutions. The third label of the third training sample may be a corresponding historical battery output power actual value. The third label may be manually labeled.

The parameters of the second embedded layer are obtained through the training mode, so that the problem that labels are difficult to obtain when the second embedded layer is independently trained can be solved, and the second embedded layer can better obtain the corresponding relation between the second embedded layer and the power predicted value.

In some embodiments, the trained first embedded layer in the fault model may be used as the second embedded layer of the power prediction model without separately training the second embedded layer.

In some embodiments, the second embedded layer of the power prediction model may share parameters with the first embedded layer of the fault model.

In some embodiments, the input to the power prediction layer may also include a battery period feature 430. In some embodiments, the inputs to the power prediction layer 640 of the power prediction model may include the operational data 410, the candidate solutions 620, the battery period characteristics 430, and the like, and the outputs may include a battery output power prediction value 650 for each of the candidate solutions.

In some embodiments, the corresponding fourth training samples may include historical work data, a plurality of historical candidate solutions, and historical battery time period characteristics, and the fourth label of the fourth training samples may be the corresponding actual predicted value of the historical battery output power. The fourth label may be manually labeled.

In some embodiments of the present disclosure, by adding a battery period feature to the input of the power prediction layer, the battery output power may be predicted based on the battery period feature, so that the predicted battery output power predicted value is more accurate.

In some embodiments, the input to the power prediction layer may also include a fault type 630. In some embodiments, the inputs of the power prediction layer 640 may include the operational data 410, the candidate solutions 620, the fault type 630, and/or the battery period characteristics 430, and the outputs may include a battery output power prediction value 650 corresponding to each candidate solution. Wherein the fault type 630 may be obtained by a fault model. Details regarding the fault model may be found in the description of the rest of the specification, for example, in fig. 4.

In some embodiments, the corresponding fifth training samples may include historical work data, historical candidate solutions, historical feature fault types, and/or historical battery period features, and the fifth label of the fifth training samples may be the corresponding historical battery output power predicted actual value. The fifth label may be manually labeled.

In some embodiments of the present disclosure, by adding a fault type to the input of the power prediction model, the predicted battery output power prediction value may be made more accurate for the fuel cell that is at the time of the fault.

In some embodiments of the present disclosure, a predicted value of the output power of the battery is determined based on the working data, the plurality of candidate solutions, the fault type and/or the battery time period characteristics, so that the prediction of the output power is more accurate by considering the influence of a plurality of factors on the output power, and further, the fuel cell can always maintain a higher output power by continuously adjusting the anode-cathode voltage value solution and the anode-cathode voltage value solution, thereby further contributing to further improving the performance and the utilization rate of the battery.

In some embodiments, a preset algorithm may be used to determine a final solution for the fuel cell based on the operational data. In some embodiments, the preset algorithm may include: generating a plurality of candidate combinations 710, wherein the candidate combinations 710 may include cathode pressure values and anode pressure values of the fuel cells; multiple rounds of iterative updates are performed on the multiple candidate combinations to determine the final solution 750. At least one of the multiple iterations includes: and updating the candidate combination based on the pressure value change amplitude corresponding to the relation update of the historical optimal combination for at least one candidate combination, wherein the historical optimal combination is determined based on an evaluation algorithm.

Candidate combinations may refer to combinations that are used to iteratively update the determination of the final solution. As shown in FIG. 7, candidate combination 710 may include candidate combination 1, candidate combination 2, candidate combination iEtc. Candidate combinations 710 may be represented by vectors. The candidate combinations may include a candidate cathode pressure value and a candidate anode pressure value for each candidate combination, i.eX _i Represents the i-th candidate combination, X _i1 Representing candidate cathode pressure values, X, in the ith candidate combination _i2 Represents the anode pressure value in the i-th candidate combination, and k represents the number of iterative updates. In the first iteration, k is 1.

In some embodiments, processor 130 may generate the plurality of candidate combinations for the first round of iterative updating in a variety of ways. Candidate combinations may be generated by random methods or acquired based on historical data.

In some embodiments, the plurality of candidate combinations may be candidate solutions for which the battery output power prediction value satisfies a set condition, wherein the set condition includes a set threshold value that is less than a preset threshold value, wherein the preset threshold value is a threshold value in the preset condition.

The set condition may refer to a condition that the predicted value of the battery output power corresponding to the candidate solution needs to be satisfied when the candidate combination is screened. In some embodiments, the set condition may include a variety, such as setting a threshold value, or the like. The set threshold may refer to a minimum value of battery output power predictions corresponding to the candidate solutions when determining the candidate combinations. The processor 130 may determine a candidate solution with a battery output power prediction value greater than a set threshold as a candidate combination. The set threshold is less than the preset threshold, and the number of candidate solutions for which the battery output power predicted value is greater than the set threshold is greater than or equal to the number of candidate solutions for which the battery output power predicted value is greater than the preset threshold. Details regarding the preset conditions and the preset threshold values can be found in the description of the other contents of the present specification, for example, fig. 5.

In some embodiments of the present disclosure, the candidate combination is determined by using the candidate solution, so that accuracy of the determined candidate combination can be ensured to a certain extent, which is beneficial to faster convergence when updating iteration, reduces iteration times, and improves iteration efficiency. By setting a threshold for the candidate solution to determine the candidate combinations and setting the threshold to be smaller than a preset threshold, the number of the determined candidate combinations can be made larger, and the final solution determined later can be made more accurate.

In some embodiments, at least one of the multiple rounds of iterative updating includes: and updating the candidate combination based on the pressure value change amplitude corresponding to the relation update of the historical optimal combination for at least one candidate combination, wherein the historical optimal combination is determined based on an evaluation algorithm. Details of the historical optimum combination and evaluation algorithm may be found in the following. As shown in fig. 7, the iterative updating of the candidate combinations may be based on the following steps.

In step 720, the magnitude of the pressure value change is updated.

The pressure value change amplitude refers to the update amplitude of the candidate cathode pressure value and the candidate anode pressure value in the candidate combination. In some embodiments, the magnitude of the pressure value change may include multiple sets. For example, the number of pressure value variation amplitudes is the same as the number of candidate combinations. Each of the plurality of sets of magnitudes of pressure value variation corresponds one-to-one with each of the candidate combinations.

In some embodiments, each set of pressure value magnitudes of variation may include a plurality of sub-magnitudes of variation. Each sub-amplitude of variation represents an updated amplitude of the pressure value (e.g., candidate cathode pressure value or candidate anode pressure value) in the candidate combination corresponding to the amplitude of variation of the pressure value. For example, the magnitude of the pressure value change corresponding to the ith candidate combination may be expressed asWherein V is _i1 、V _i2 Respectively representing sub-variation amplitudes corresponding to the candidate cathode pressure value and the candidate anode pressure value in the ith candidate combination, and k represents the iteration times. In some embodiments, the initial values of the magnitudes of pressure value changes corresponding to the plurality of candidate combinations may be the same or different. Wherein the initial value of the amplitude of the pressure value variation may be generated on a random basis.

In some embodiments, a history optimal combination may be used to determine the magnitude of the pressure value change. The processor may update the magnitude of the pressure value change based on the relationship of the candidate combination to the historical optimal combination. For example, if the difference between the candidate combination and the historical optimal combination is small, the corresponding pressure value variation amplitude is small; and vice versa, larger.

In some embodiments, the historical optimal combination may include an independent optimal combination corresponding to the candidate combination, and an associated optimal combination that collectively corresponds to the plurality of candidate combinations.

In some embodiments, for each of the plurality of candidate combinations, the historical optimal combination includes an independent optimal combination corresponding to each candidate combination, and an associated optimal combination that corresponds in common with the plurality of candidate combinations. Wherein, the associated optimal combinations corresponding to the candidate combinations are the same, and the independent optimal combinations are different. In some embodiments, the historical optimal combination may be determined based on an evaluation algorithm. For a detailed description of the evaluation algorithm, reference is made to the following relevant text.

The historical optimal combination refers to a candidate combination with the maximum predicted value of the battery output power in the historical iterative updating process. For more description of battery output power predictions, see the relevant description elsewhere.

The independent optimal combination corresponding to the ith candidate combination refers to the updating candidate combination with the largest battery output power predicted value corresponding to a plurality of updating candidate combinations corresponding to the ith candidate combination by the current iteration updating wheel. For example, in the k-th iteration, the independent optimal combination corresponding to the i-th candidate combination may be the updated candidate combination with the largest battery output power predicted value corresponding to all updated i-th candidate combinations in the previous k-1 iteration.

The associated optimal combination corresponding to the ith candidate combination is the updated candidate combination with the largest battery output power predicted value corresponding to all updated candidate combinations corresponding to the plurality of candidate combinations up to the current iteration round. For example, in the kth iteration, the associated optimal combination may be the updated candidate combination with the highest predicted battery output power value during the previous k-1 iterations.

By combining independent and associated optimal combinations in some embodiments, the exploration process may be made to better combine local exploration with global scenarios.

In some embodiments, the historical optimal combination may be determined based on an evaluation algorithm. In some embodiments, the evaluation algorithm may include: determining a battery output power predicted value corresponding to each candidate combination in the plurality of candidate combinations through a power prediction model based on the plurality of candidate combinations and the working data; based on the battery output power predictions, a historical optimal combination is determined. Details regarding the power prediction model may be found in the description of the other contents of this specification, for example, fig. 6.

In some embodiments, the evaluation algorithm may refer to an algorithm that evaluates candidate combinations based on a machine learning model, preset rules, or the like. In some embodiments, the evaluation algorithm may include predicting iterating battery output power predictions for each candidate combination through a power prediction model based on the candidate combinations, and determining a historical optimal combination based on the battery output power predictions. Specifically, based on each of a plurality of candidate combinations and corresponding working data, the power prediction model is input to obtain corresponding battery output power predicted values, the battery output power predicted value in a plurality of iterations in each candidate combination is the independent optimal combination, and the battery output power predicted value in the plurality of iterations in all candidate combinations is the associated optimal combination.

By some embodiments of the present description, a historical optimal combination may be more accurately determined based on a battery output power prediction value, and further improve iteration efficiency.

In some embodiments, the processor updating the magnitudes of pressure value changes corresponding to the plurality of candidate combinations may refer to determining, for each magnitude of pressure value change, a sub-magnitude of change based on the following formula: updated sub-variance = weight 1 original candidate combination sub-variance + weight 2 first difference + weight 3 second difference. The updated sub-variation amplitude may be the sub-variation amplitude corresponding to the candidate cathode pressure value or the candidate anode pressure value in the candidate combination corresponding to the next round of updating iteration. The sub-variation amplitude of the original candidate combination can be the sub-variation amplitude corresponding to the candidate cathode pressure value or the candidate anode pressure value in the candidate combination corresponding to the current round of updating iteration. The first difference value corresponds to the difference value between the candidate combination and the independent optimal combination; the second difference value corresponds to a difference value of the candidate combination and the associated optimal combination. The weights 1, 2 and 3 may be preset, or may be determined by other manners, for example, determined based on an algorithm such as regression analysis.

Step 730, updating the candidate combinations.

At iteration 1, the plurality of candidate combinations includes a plurality of initial combinations. As shown in fig. 7, the plurality of initial combinations includes initial combination 1, initial combination 2, initial combination i, and the like.

The iterative updating of the candidate combinations by the processor includes iteratively updating each candidate combination based on the sub-variation magnitudes corresponding to each candidate combination. For example, the sub-variation amplitude may be added to the original candidate combination to obtain an updated candidate combination, i.e., the updated candidate combination may be expressed asWherein (1)>Can be expressed as +.>

By the method, the search can be performed in different directions in the solution space based on a plurality of groups of different candidate combinations. The direction of exploration and the magnitude of the variation amplitude can be dynamically adjusted according to the comparison with the historical optimal combination, so that exploration is more targeted and the final solution is approached more quickly.

Step 740, iterate whether the termination condition is satisfied.

In some embodiments, when the iteration meets a preset termination condition when the iteration candidate combination is performed, the iteration update is ended. The termination condition may be that the iteration number reaches an iteration number threshold, the predicted value of the battery output power corresponding to the candidate combination reaches a preset value, and the like, and the iteration is stopped. For example, the iteration number threshold is set to 100 times, and when the iteration number reaches 100 times, the iteration is stopped. And setting a preset value, and stopping iteration when the predicted value of the battery output power reaches the preset value.

In some embodiments, after each iteration update is completed, at least one of the sets of candidate combinations is determined whether it meets an iteration termination condition. If the iteration satisfies the iteration termination condition, the iteration is ended, and the historical optimal combination is determined as a final solution 750. And if the iteration does not meet the termination condition, continuing the next iteration until the iteration termination condition is met.

In some embodiments of the present description, exploration is performed in the solution space in different directions based on a plurality of candidate combinations. According to the comparison of the candidate combinations and the historical optimal combination, the direction of exploration and the magnitude of the variation amplitude of the pressure value can be dynamically adjusted, so that exploration is more targeted and the final solution is approached more quickly. And updating the candidate combinations in an iterative mode, and optimizing the candidate combinations. Therefore, the final solution with the maximum fuel cell output power is determined, and the performance and the utilization rate of the fuel cell are further improved while energy is saved.

It should be noted that the above description of the flow is only for the purpose of illustration and description, and does not limit the application scope of the present specification. Various modifications and changes to the flow may be made by those skilled in the art under the guidance of this specification. However, such modifications and variations are still within the scope of the present description.

While the basic concepts have been described above, it will be apparent to those skilled in the art that the foregoing detailed disclosure is by way of example only and is not intended to be limiting. Although not explicitly described herein, various modifications, improvements, and adaptations to the present disclosure may occur to one skilled in the art. Such modifications, improvements, and modifications are intended to be suggested within this specification, and therefore, such modifications, improvements, and modifications are intended to be included within the spirit and scope of the exemplary embodiments of the present invention.

Meanwhile, the specification uses specific words to describe the embodiments of the specification. Reference to "one embodiment," "an embodiment," and/or "some embodiments" means that a particular feature, structure, or characteristic is associated with at least one embodiment of the present description. Thus, it should be emphasized and should be appreciated that two or more references to "an embodiment" or "one embodiment" or "an alternative embodiment" in various positions in this specification are not necessarily referring to the same embodiment. Furthermore, certain features, structures, or characteristics of one or more embodiments of the present description may be combined as suitable.

Likewise, it should be noted that in order to simplify the presentation disclosed in this specification and thereby aid in understanding one or more inventive embodiments, various features are sometimes grouped together in a single embodiment, figure, or description thereof. This method of disclosure, however, is not intended to imply that more features than are presented in the claims are required for the present description. Indeed, less than all of the features of a single embodiment disclosed above.

In some embodiments, numbers describing the components, number of attributes are used, it being understood that such numbers being used in the description of embodiments are modified in some examples by the modifier "about," approximately, "or" substantially. Unless otherwise indicated, "about," "approximately," or "substantially" indicate that the number allows for a 20% variation. Accordingly, in some embodiments, numerical parameters set forth in the specification and claims are approximations that may vary depending upon the desired properties sought to be obtained by the individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and employ a method for preserving the general number of digits. Although the numerical ranges and parameters set forth herein are approximations that may be employed in some embodiments to confirm the breadth of the range, in particular embodiments, the setting of such numerical values is as precise as possible.

Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments of this specification. Other variations are possible within the scope of this description. Thus, by way of example, and not limitation, alternative configurations of embodiments of the present specification may be considered as consistent with the teachings of the present specification. Accordingly, the embodiments of the present specification are not limited to only the embodiments explicitly described and depicted in the present specification.

Claims

1. A method of water management for a fuel cell, the method comprising:

acquiring working data of a fuel cell;

determining a normal condition or fault type and a final solution of the fuel cell based on the working data, wherein the final solution comprises a cathode pressure value solution and an anode pressure value solution of the fuel cell, the cathode pressure value solution is a pressure value corresponding to a cathode of the fuel cell when a preset condition is met, and the anode pressure value solution is a pressure value corresponding to an anode of the fuel cell when the preset condition is met;

said determining a normal condition or a fault type of said fuel cell based on said operational data comprises:

determining a battery time period characteristic based on the operation data, the battery time period characteristic being characteristic of operation data of the fuel cell at a plurality of different time points of a period of time preceding a current time point;

the battery time period characteristics at least comprise a voltage maximum value, a current maximum value, a temperature maximum value, a maximum voltage change rate and a maximum temperature change rate of a period of time before the current time point;

determining the normal condition or the fault type based on the battery period characteristics;

The fault type is the type of fault cause causing the battery to fail to operate normally, and at least comprises flooding, film drying and other types;

said determining a final solution for said fuel cell based on said operational data comprises:

determining the final solution through a preset algorithm based on the working data;

said determining the final solution comprises:

generating a plurality of candidate solutions, the candidate solutions including the cathode pressure value solution and the anode pressure value solution of the fuel cell;

performing multiple rounds of iterative updating on the multiple candidate solutions to determine the final solution;

at least one of the rounds of iterations includes: updating, for at least one of the candidate solutions, an adjustment increment corresponding to a relation update of a historical final solution based on the adjustment increment, the candidate solution being updated based on the adjustment increment; wherein the historical final solution is determined based on an evaluation algorithm; determining an adjustment mode based on the normal condition or the fault type;

the regulation modes are different modes for regulating the water content in the battery according to different running states of the battery; the regulating mode at least comprises a normal regulating mode, a flooding regulating mode and a membrane dry regulating mode;

Judging that the battery has a flooding fault in response to the fact that the water content in the battery exceeds a water content maximum threshold or the water saturation of the gas diffusion layer exceeds a saturation maximum threshold, and determining that the regulation mode is the flooding regulation mode;

judging that the membrane dry fault occurs in the battery in response to the fact that the water content in the battery is lower than a water content minimum threshold value or the water saturation of the gas diffusion layer is lower than a saturation minimum threshold value, and determining the regulating mode to be the membrane dry regulating mode;

and adjusting a cathode pressure value of the fuel cell to the cathode pressure value solution and adjusting an anode pressure value of the fuel cell to the anode pressure value solution based on the adjustment mode.

2. The method of claim 1, wherein the determining a final solution for the fuel cell based on the operational data comprises:

determining the plurality of candidate solutions based on the working data;

determining a battery output power predicted value corresponding to each candidate solution in the plurality of candidate solutions through a power prediction model based on the plurality of candidate solutions and the working data, wherein the power prediction model is a machine learning model;

and determining the final solution based on the battery output power predicted value, wherein the final solution is the corresponding candidate solution of which the battery output power predicted value meets the preset condition.

3. A water management system for a fuel cell, the system comprising:

the acquisition module is used for acquiring the working data of the fuel cell;

a first determining module, configured to determine, based on the operation data, a normal condition or a fault type of the fuel cell and a final solution, where the final solution includes a cathode pressure value solution and an anode pressure value solution of the fuel cell, where the cathode pressure value solution is a pressure value corresponding to a cathode of the fuel cell when a preset condition is met, and the anode pressure value solution is a pressure value corresponding to an anode of the fuel cell when the preset condition is met;

the first determination module is further to:

said determining the final solution comprises:

at least one of the rounds of iterations includes: updating, for at least one of the candidate solutions, an adjustment increment corresponding to a relation update of a historical final solution based on the adjustment increment, the candidate solution being updated based on the adjustment increment; wherein the historical final solution is determined based on an evaluation algorithm;

a second determining module for determining an adjustment mode based on the normal condition or the fault type;

and the adjusting module is used for adjusting the cathode pressure value of the fuel cell to the cathode pressure value solution and adjusting the anode pressure value of the fuel cell to the anode pressure value solution based on the adjusting mode.

4. The system of claim 3, wherein the first determination module is further to:

determining the plurality of candidate solutions based on the working data;

5. A water management device for a fuel cell, said device comprising at least one processor and at least one memory;

the at least one memory is configured to store computer instructions;

the at least one processor is configured to execute at least some of the computer instructions to implement the method of any one of claims 1 to 2.

6. A computer readable storage medium storing computer instructions which, when executed by a processor, implement the method of any one of claims 1 to 2.