A fuel cell flooding control method, device and storage medium

Abstract

The application discloses a fuel cell water flooding control method, device and storage medium. By constructing a dynamic feature set and calculating a real-time entropy weight score, the water flooding risk state of the anode of the fuel cell can be comprehensively and accurately reflected, and the early warning capability and diagnostic accuracy of the water flooding fault are improved. Based on the real-time entropy weight score, the output power, the loading rate and the drainage flow are dynamically adjusted to realize adaptive control of the water management of the fuel cell, and the hysteresis and excessive regulation problems of the traditional fixed threshold control strategy are avoided. The dependence on artificial experience is reduced, the water flooding risk grade is automatically identified through the data-driven entropy weight analysis method, the water flooding fault probability caused by subjective judgment errors is reduced, the reliability and durability of the fuel cell system are enhanced, and the service life of the fuel cell stack is effectively prolonged.

Description

Technical Field

[0001] The present invention relates to the technical field of fuel cells, and in particular to a fuel cell flooding control method, device and storage medium. Background Art

[0002] With the rapid development of new energy technologies, fuel cells, with their high efficiency and environmentally friendly advantages, have been widely used in fields such as automobiles and distributed power generation. However, during fuel cell operation, anode flooding seriously affects their performance and lifespan. Flooding can hinder gas transmission, reduce electrochemical reaction efficiency, and even cause system failure. Therefore, effective control of fuel cell anode flooding is critical to ensuring stable operation.

[0003] Existing methods for mitigating fuel cell anode flooding mostly mitigate this problem through adjustments within the fuel cell system itself (such as increasing hydrogen flow and adjusting the exhaust valve opening). However, in practical applications such as electric vehicles, fuel cell systems typically work in conjunction with lithium-ion batteries, with an energy management system coordinating the output power of the fuel cell system and the lithium-ion batteries.

[0004] When fuel cell anode flooding occurs, if the energy management system increases the output power of the fuel cell system, the fuel cell anode flooding will be exacerbated. Therefore, there is an urgent need for a method that can prevent the anode flooding of the fuel cell from being exacerbated by coordinating the fuel cell system and the energy management system in the event of fuel cell anode flooding. Summary of the Invention

[0005] In response to the above-mentioned problems, the present invention proposes a fuel cell flooding control method, device and storage medium. By constructing a dynamic feature set and calculating a real-time entropy weight score, accurate diagnosis and adaptive control of the fuel cell anode flooding risk can be achieved, effectively solving technical problems existing in traditional methods such as diagnostic delay, poor control adaptability, insufficient system robustness and the impact of fuel cell durability on flooding.

[0006] In a first aspect, an embodiment of the present disclosure provides a fuel cell flooding control method, the method comprising:

[0007] Obtaining a feature vector of the fuel cell, and inputting the feature vectors into a dynamically updated sliding window one by one to form a dynamic feature set containing features of the most recent N time windows, where N is a preset storage capacity of the sliding window;

[0008] Calculate the real-time entropy weight score corresponding to the dynamic feature set, and determine the flooding risk level of the fuel cell anode based on the real-time entropy weight score;

[0009] When the real-time entropy weight score is greater than a first preset threshold, dynamically adjusting the output power and loading rate of the fuel cell according to the real-time entropy weight score;

[0010] The drainage flow of the fuel cell drain valve is dynamically adjusted according to the flooding risk level.

[0011] Optionally, obtaining a characteristic vector of the fuel cell includes: collecting the inlet and outlet air pressures and the stack voltage of the fuel cell anode; calculating the inlet and outlet air pressure difference of the anode based on the inlet and outlet air pressures, and determining the air pressure difference change rate of the anode based on the inlet and outlet air pressure difference; determining the voltage fluctuation value based on the stack voltage; and forming a characteristic vector based on the inlet and outlet air pressure difference, the air pressure difference change rate and the voltage fluctuation value.

[0012] Optionally, the real-time entropy weight score corresponding to the dynamic feature set is calculated, including: normalizing the range of each window feature in the dynamic feature set to generate a standard feature; taking each standard feature as the target feature in turn, calculating the feature normalization ratio based on the target feature, and calculating the information entropy according to the feature normalization ratio; determining the feature weight of the target feature according to the information entropy calculation; and performing weighted summation of each target feature and its corresponding feature weight through an entropy weight model to determine the real-time entropy weight score.

[0013] Optionally, the method also includes: obtaining historical entropy weight scores, calculating the score deviation between the real-time entropy weight scores and the historical entropy weight scores; obtaining the historical distribution corresponding to each standard feature, and calculating the individual deviation corresponding to each standard feature based on the historical distribution; when the score deviation is greater than a preset score deviation threshold and each individual deviation is within a preset normal range, determining that the entropy weight model has failed; when one of the individual deviations is greater than a preset high abnormality threshold, determining that the sensor has failed; when multiple individual deviations are greater than the preset abnormality threshold, determining that the fuel cell system level has failed.

[0014] Optionally, the flooding risk level of the fuel cell anode is determined based on the real-time entropy weight score, including: when the real-time entropy weight score is less than or equal to a first preset threshold, determining the flooding risk level as normal; when the real-time entropy weight score is greater than the first preset threshold and less than or equal to the second preset threshold, determining the flooding risk level as mild flooding; when the real-time entropy weight score is greater than the second preset threshold and less than or equal to the third preset threshold, determining the flooding risk level as moderate flooding; when the real-time entropy weight score is greater than the third preset threshold, determining the flooding risk level as severe flooding.

[0015] Optionally, the output power and loading rate of the fuel cell are dynamically adjusted according to the real-time entropy weight score, including: obtaining the basic weight and weight gain coefficient of the lithium battery, calculating the product of the real-time entropy weight score and the weight gain coefficient, and adding the product to the basic weight to obtain the priority weight of the lithium battery; determining the target weight of the fuel cell according to the priority weight of the lithium battery; obtaining the total load power, and allocating the output power of the fuel cell according to the product of the target weight and the total load power; obtaining the anode gas pressure difference change rate, and calculating the loading rate of the fuel cell according to the anode gas pressure difference change rate and the real-time entropy weight score; wherein the loading rate is calculated using the following formula:

[0016] ;

[0017] in, is the loading rate, is the preset maximum loading rate, is the preset safety threshold pressure difference change rate, Score the real-time entropy weight, is the anode gas pressure difference change rate, and t is the current moment.

[0018] Optionally, the drainage flow of the fuel cell drain valve is dynamically adjusted according to the flooding risk level, including: determining a basic drainage coefficient corresponding to the flooding risk level, obtaining a target change in real-time state parameters, wherein the real-time state parameters include anode gas pressure difference change rate and a stack voltage fluctuation amplitude change rate; obtaining fuzzy inference rules, wherein the fuzzy inference rules include drainage correction amounts corresponding to each change amount; correcting the basic drainage coefficient by the drainage correction amount to generate a target drainage coefficient, and determining the drainage flow of the fuel cell drain valve according to the target drainage coefficient.

[0019] Optionally, after dynamically adjusting the drainage flow of the fuel cell drain valve according to the flooding risk level, the method further includes: determining the duration when the real-time entropy weight score is less than a first preset threshold and the anode gas pressure difference change rate and the stack voltage fluctuation amplitude are both less than the corresponding thresholds; and releasing the flooding mode when the duration is greater than the preset time threshold.

[0020] In a second aspect, an embodiment of the present disclosure further provides an electronic device, the electronic device comprising:

[0021] At least one processor; and a memory in communication with the at least one processor; wherein, when the memory stores a computer program executable by the at least one processor, the computer program is executed by the at least one processor so that the at least one processor can execute a fuel cell flooding control method as described in any embodiment of the present disclosure.

[0022] In a third aspect, an embodiment of the present disclosure provides a computer storage medium having a computer program stored thereon, which, when executed by a processor, implements a fuel cell flooding control method as described in any embodiment of the present disclosure.

[0023] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the present invention will become readily understood through the following description.

[0024] Therefore, the present invention has the following beneficial effects:

[0025] 1. By constructing a dynamic feature set and calculating the real-time entropy weight score, the water flooding risk status of the fuel cell anode can be comprehensively and accurately reflected, improving the early warning capability and diagnostic accuracy of water flooding failures.

[0026] 2. Dynamically adjust the output power, loading rate and drainage flow based on the real-time entropy weight score, realizing adaptive control of fuel cell water management and avoiding the hysteresis and over-regulation problems of traditional fixed threshold control strategies.

[0027] 3. Reduced reliance on manual experience, automatic identification of flooding risk levels through data-driven entropy weight analysis methods, reduced the probability of flooding failures caused by subjective misjudgment, enhanced the reliability and durability of the fuel cell system, and effectively extended the life of the stack. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 This is a flow chart of a fuel cell flooding control method provided according to the first embodiment of the present invention;

[0029] Figure 2 is a flow chart of another fuel cell flooding control method provided according to the second embodiment of the present invention;

[0030] Figure 3 This is a schematic structural diagram of a fuel cell flooding control device provided according to a third embodiment of the present invention;

[0031] Figure 4 It is a structural diagram of an electronic device provided according to the fourth embodiment of the present invention. DETAILED DESCRIPTION

[0032] To make the objectives, technical solutions, and advantages of the embodiments of the present application more clear, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all of the embodiments. Generally, the components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations.

[0033] In the description of this application, it should be noted that the terms "inner" and "outer" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, or the orientations or positional relationships in which the product of this application is typically placed when in use. These terms are intended solely to facilitate the description of this application and simplify the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limitations on this application. Furthermore, the terms "first" and "second" and the like are used solely for distinction and should not be construed as indicating or implying relative importance.

[0034] It should also be noted that, in the description of this application, unless otherwise expressly specified or limited, the terms "disposed" and "connected" should be understood in a broad sense. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to direct connections, indirect connections through an intermediate medium, or internal connections between two components. Those skilled in the art will understand the specific meanings of the above terms in this application based on the specific circumstances.

[0035] The present invention will be described in detail below with reference to the accompanying drawings.

[0036] Example 1

[0037] Figure 1 A flowchart of a fuel cell flooding control method is provided for Embodiment 1 of the present invention. This embodiment is applicable to fuel cell scenarios. The method can be executed by a fuel cell flooding control device provided in an embodiment of the present disclosure. The device can be implemented using software and / or hardware and can generally be integrated into a computer device. The method of the embodiment of the present disclosure specifically includes:

[0038] S110: Acquire a feature vector of the fuel cell, and input the feature vectors into a dynamically updated sliding window one by one to form a dynamic feature set containing features of the windows at the latest N moments, where N is a preset storage capacity of the sliding window.

[0039] The eigenvector is a key parameter combination used to characterize the operating status of a fuel cell and serves as the core basis for determining the risk of anode flooding. The eigenvector is composed of the anode inlet and outlet pressure differential, the rate of change of the pressure differential, and the stack voltage fluctuation. By integrating multiple parameters, it can more comprehensively reflect the potential risk of anode flooding. The dynamically updated sliding window retains only the feature data from the most recent N moments for real-time analysis and model parameter updates. N is the preset storage capacity of the sliding window, representing the maximum amount of data the window can accommodate. This can be 500, meaning the sliding window stores the most recent 500 frames of feature data. When new data is added to the end of the window, if the data volume exceeds N, the oldest data entry is automatically discarded to ensure the window always contains the latest data. The dynamic feature set refers to the set of feature vectors from the most recent N moments stored within the sliding window, which is continuously updated over time.

[0040] Optionally, obtaining a characteristic vector of the fuel cell includes: collecting the inlet and outlet air pressures and the stack voltage of the fuel cell anode; calculating the inlet and outlet air pressure difference of the anode based on the inlet and outlet air pressures, and determining the air pressure difference change rate of the anode based on the inlet and outlet air pressure difference; determining the voltage fluctuation value based on the stack voltage; and forming a characteristic vector based on the inlet and outlet air pressure difference, the air pressure difference change rate and the voltage fluctuation value.

[0041] Specifically, the fuel cell anode inlet and outlet pressures and the stack voltage are fundamental parameters reflecting the fuel cell's operating status. The controller calculates the anode inlet and outlet pressure differential based on the collected pressures. This is done by subtracting the anode outlet pressure from the anode inlet pressure. This pressure differential reflects the flow resistance of the anode gas channel. If flooding occurs, the pressure differential will exhibit abnormal fluctuations. The controller is a computer-based controller that manages fuel cell flooding. The controller then determines the anode pressure differential change rate based on the inlet and outlet pressure differential. This is done by calculating the pressure differential change per unit time in real time. Analysis of the pressure differential change rate can be used to determine the speed of flooding. Furthermore, the controller determines the voltage fluctuation value based on the stack voltage. This voltage fluctuation value reflects the stability of the stack output voltage. Flooding can cause abnormal stack voltage fluctuations. Finally, the three parameters—the collected inlet and outlet pressure differential, the calculated pressure differential change rate, and the voltage fluctuation value—are combined to form a feature vector used to assess the risk of fuel cell anode flooding.

[0042] S120: Calculate the real-time entropy weight score corresponding to the dynamic feature set, and determine the flooding risk level of the fuel cell anode according to the real-time entropy weight score.

[0043] The real-time entropy weight score is a comprehensive risk score calculated using the entropy weight method based on the dynamic feature set within a sliding window. It is used to quantify the risk of anode flooding. The score ranges from [0 to 1], with larger values ​​indicating a higher flooding risk. The flooding risk level is a risk classification based on the real-time entropy weight score and is used to guide the execution of subsequent control strategies. Different levels correspond to different power allocation strategies and drainage intensities, achieving a graded response.

[0044] Optionally, the real-time entropy weight score corresponding to the dynamic feature set is calculated, including: normalizing the range of each window feature in the dynamic feature set to generate a standard feature; taking each standard feature as the target feature in turn, calculating the feature normalization ratio based on the target feature, and calculating the information entropy according to the feature normalization ratio; determining the feature weight of the target feature according to the information entropy calculation; and performing weighted summation of each target feature and its corresponding feature weight through an entropy weight model to determine the real-time entropy weight score.

[0045] Specifically, the controller will perform range normalization on each window feature in the dynamic feature set to eliminate the dimensional effects of different parameters. Range normalization is divided into two processing methods: positive indicators and negative indicators. For positive indicators, such as air pressure difference, a larger value may indicate a higher flood risk. The calculation formula for the range normalization process is:

[0046] ;

[0047] in, is the value of the j-th feature in the i-th sample, and is the maximum and minimum value of the j-th feature in the current window, is the normalized value of the jth feature in the i-th sample. For negative indicators, such as voltage fluctuation, a larger value may indicate a higher risk, but the indicator is negative in nature. The calculation formula for the range normalization process is:

[0048] ;

[0049] in, is the value of the j-th feature in the i-th sample, and is the maximum and minimum value of the j-th feature in the current window, is the normalized value of the j-th feature in the i-th sample.

[0050] Furthermore, the controller will take each standard feature as the target feature in turn and calculate its feature normalization ratio. For the jth feature, the standard feature value z of N samples in the window is first calculated. ij Perform normalization and calculate the normalized ratio of the i-th sample in the j-th feature:

[0051] ;

[0052] Where N is the number of samples in the window, is the standardized value of the j-th feature in the i-th sample, is the normalized ratio of the i-th sample under the j-th feature. The normalized ratio reflects the relative importance of a single sample in the feature. Then, the controller can calculate the information entropy based on the normalized ratio of the feature. The calculation formula of information entropy is:

[0053] ;

[0054] in, is a constant used to standardize the entropy range, represents the information entropy of the jth feature, N is the number of samples in the window, is the normalized ratio of the i-th sample under the j-th feature. Information entropy measures the uncertainty of a feature. A smaller value indicates more ordered information about the feature and a greater contribution to risk assessment.

[0055] Then, the controller calculates the feature weight of the target feature based on the information entropy. The calculation formula of the feature weight is:

[0056] ;

[0057] Among them, m is the dimension of the feature vector. For example, when the feature vector contains 3 parameters, m=3. represents the information entropy of the j-th feature, represents the feature weight of the jth feature. This formula is normalized so that the sum of the feature weights is 1. The larger the weight, the more important the feature is in flood risk assessment.

[0058] Finally, the controller uses the entropy weight model to perform a weighted summation of each target feature and its corresponding feature weight to obtain a real-time entropy weight score. The formula is:

[0059] ;

[0060] in, is the jth feature normalization value of the current new data point, is the real-time entropy weight score at time t, m is the number of dimensions of the feature vector, The real-time entropy weight score integrates the dynamic weights of multiple features and can quantify the flooding risk level of the fuel cell anode, providing a basis for subsequent control strategies.

[0061] Optionally, the flooding risk level of the fuel cell anode is determined based on the real-time entropy weight score, including: when the real-time entropy weight score is less than or equal to a first preset threshold, determining the flooding risk level as normal; when the real-time entropy weight score is greater than the first preset threshold and less than or equal to the second preset threshold, determining the flooding risk level as mild flooding; when the real-time entropy weight score is greater than the second preset threshold and less than or equal to the third preset threshold, determining the flooding risk level as moderate flooding; when the real-time entropy weight score is greater than the third preset threshold, determining the flooding risk level as severe flooding.

[0062] Specifically, three preset thresholds can be set as critical values ​​for risk level division, where the first preset threshold is the lowest critical value for judging whether flooding has occurred, and the second preset threshold and the third preset threshold are used to distinguish flooding levels of different severity. After obtaining the real-time entropy weight score, the controller will compare it with each preset threshold: if the real-time entropy weight score is less than or equal to the first preset threshold, it can be set to 0.3, indicating that the parameters such as the gas pressure difference, gas pressure difference change rate and voltage fluctuation of the fuel cell anode are all within the normal range and there is no sign of flooding, so the flooding risk level is determined to be normal; if the real-time entropy weight score is greater than the first preset threshold and less than or equal to the second preset threshold, it can be set to 0.6, indicating that the anode has a certain degree of flooding characteristics, such as a slight increase in gas pressure difference or slightly abnormal voltage fluctuation, but the overall risk is still acceptable. If the real-time entropy weight score is greater than the second preset threshold and less than or equal to the third preset threshold, which can be set to 0.8, it indicates that the flooding situation is relatively obvious, the air pressure difference and the air pressure difference change rate continue to increase, the voltage fluctuation amplitude increases, and it has a significant impact on the performance of the fuel cell. Therefore, the flooding risk level is determined to be moderate. If the real-time entropy weight score is greater than the third preset threshold, it is determined that the anode flooding is very serious, the gas channel may be close to being blocked, the fuel cell stack performance has dropped significantly, and there is a risk of shutdown. At this time, the flooding risk level is severe. Through the graded determination method, corresponding control strategies can be accurately triggered for different degrees of flooding risk, such as dynamically adjusting the power allocation weight and optimizing the drainage flow, to improve the stability and risk resistance of the fuel cell system.

[0063] S130: When the real-time entropy weight score is greater than a first preset threshold, dynamically adjusting the output power and loading rate of the fuel cell according to the real-time entropy weight score.

[0064] The first preset threshold is the critical score that triggers fault-tolerant control, which can be 0.3. The loading rate is the rate of change of the fuel cell output power, measured in kW / s, reflecting the speed of power adjustment. The loading rate is dynamically reduced based on flooding risk to avoid exacerbating flooding due to sudden power changes, such as rapid power increases.

[0065] Optionally, the output power and loading rate of the fuel cell are dynamically adjusted according to the real-time entropy weight score, including: obtaining the basic weight and weight gain coefficient of the lithium battery, calculating the product of the real-time entropy weight score and the weight gain coefficient, and adding the product to the basic weight to obtain the priority weight of the lithium battery; determining the target weight of the fuel cell according to the priority weight of the lithium battery; obtaining the total load power, and allocating the output power of the fuel cell according to the product of the target weight and the total load power; obtaining the anode gas pressure difference change rate, and calculating the loading rate of the fuel cell according to the anode gas pressure difference change rate and the real-time entropy weight score; wherein the loading rate is calculated using the following formula:

[0066] ;

[0067] in, is the loading rate, is the preset maximum loading rate, is the preset safety threshold pressure difference change rate, Score the real-time entropy weight, is the anode gas pressure difference change rate, and t is the current moment.

[0068] Specifically, the controller can obtain the basic weight and weight gain coefficient of the lithium battery, where the basic weight is the system's default lithium battery power allocation ratio, which can be taken as 0.3, and the weight gain coefficient is used to amplify the regulatory effect of the real-time entropy weight score on the weight, which can be taken as 0.7. Next, the product of the real-time entropy weight score and the weight gain coefficient is calculated. This product reflects the incremental adjustment of the lithium battery weight by the current flooding risk. Add this product to the basic weight to obtain the priority weight of the lithium battery, that is, the proportion of the total load power that the lithium battery should bear. Subsequently, the target weight of the fuel cell is determined based on the lithium battery priority weight. Since the total weight is 1, the target weight of the fuel cell is 1-priority weight, which represents the proportion of load power that the fuel cell needs to bear.

[0069] In addition, when dynamically allocating power between lithium-ion batteries and fuel cells, constraints on the lithium-ion battery state of charge (SOC) and fuel cell thermal management can be added. These constraints include lithium-ion battery SOC protection and fuel cell thermal management. Lithium-ion battery SOC protection involves setting a lower SOC limit, such as 20%. When the lithium-ion battery output power causes the SOC to fall below the lower limit, the fuel cell power contribution is forcibly increased to prevent excessive discharge of the lithium-ion battery. Fuel cell thermal management involves limiting the maximum power output of the fuel cell when its temperature exceeds a safety threshold, such as 80°C, prioritizing the lithium-ion battery to bear the load and prevent overheating and flooding.

[0070] Furthermore, the controller obtains the total load power and calculates the target output power of the fuel cell based on the product of the target weight of the fuel cell and the total load power. Finally, the anode gas pressure difference change rate is obtained and the loading rate is calculated using the following formula:

[0071] ;

[0072] in, is the loading rate, is the preset maximum loading rate, is the preset safety threshold pressure difference change rate, Score the real-time entropy weight, is the rate of change of the anode gas pressure difference, and t is the current moment. This formula dynamically suppresses the fuel cell loading rate by introducing a real-time entropy weight score and the rate of change of the gas pressure difference: when the risk score is higher or the rate of change of the gas pressure difference is greater, the loading rate is limited to a lower level, thereby avoiding the deterioration of flooding caused by rapid power increase and realizing intelligent regulation of fuel cell output power changes.

[0073] S140: Dynamically adjust the drainage flow of the fuel cell drain valve according to the flooding risk level.

[0074] The drainage flow rate refers to the real-time drainage flow of the anode drain valve, which is controlled by adjusting the valve opening. The drainage intensity is adaptively adjusted according to the severity of flooding and recovery trends to avoid insufficient or excessive drainage.

[0075] Optionally, the drainage flow of the fuel cell drain valve is dynamically adjusted according to the flooding risk level, including: determining a basic drainage coefficient corresponding to the flooding risk level, obtaining a target change in real-time state parameters, wherein the real-time state parameters include anode gas pressure difference change rate and a stack voltage fluctuation amplitude change rate; obtaining fuzzy inference rules, wherein the fuzzy inference rules include drainage correction amounts corresponding to each change amount; correcting the basic drainage coefficient by the drainage correction amount to generate a target drainage coefficient, and determining the drainage flow of the fuel cell drain valve according to the target drainage coefficient.

[0076] Specifically, based on the determined flooding risk level, the controller can use a predefined basic drainage coefficient as a baseline adjustment value for the drainage flow rate. For example, a basic coefficient of 0.2 corresponds to mild flooding, 0.4 to moderate flooding, and 0.6 to severe flooding. The basic drainage coefficient is directly linked to the risk level, ensuring that drainage efforts are increased as the risk increases. Subsequently, the controller obtains target changes in real-time state parameters, including the rate of change of the anode gas pressure difference and the rate of change of the stack voltage fluctuation amplitude. These two parameters reflect the flooding development trend and the rate of stack performance degradation, respectively, and are key inputs for dynamically adjusting the drainage flow rate. A fuzzy inference rule base pre-establishes a mapping between these changes and drainage corrections. Rules employ an IF-THEN structure, for example: IF the gas pressure difference change rate is "rapidly rising" AND the voltage fluctuation change rate is "significantly increasing" THEN the drainage correction is "+0.15." Each rule corresponds to a specific state combination and outputs a corresponding correction coefficient. The corresponding correction is matched in the fuzzy rule base using the real-time input gas pressure difference change rate and voltage fluctuation change rate. For example, if the pressure differential rate of change exceeds a threshold and voltage fluctuations increase, the system triggers a higher correction factor. The correction factors are combined through weighted summation or the maximum membership principle to generate a drainage correction factor. The base drainage coefficient and the correction factor are superimposed to obtain the target drainage coefficient. The target drainage coefficient is used to adjust the baseline flow rate of the drainage valve and is expressed as follows:

[0077] ;

[0078] in, For normal drainage flow, is the adjusted drainage flow rate, and k is the correction coefficient. This dual adjustment mechanism ensures a basic response to the risk level while also enabling fine-tuning of real-time state changes through fuzzy reasoning. For example, if the rate of change of the air pressure difference suddenly increases but the risk level has not yet jumped, the system can still increase drainage in a timely manner using the correction coefficient to prevent further flooding.

[0079] Furthermore, when regulating drainage flow, the controller collects feedback signals from the drain valve opening in real time and compares them with the control command. If the deviation exceeds a threshold, such as 10%, it is determined to be a valve failure. A backup drain valve can be switched or drainage time can be increased for compensation. If the deviation exceeds 20%, a system alarm is triggered, forcing the fuel cell power to idle, reducing water production. This drainage feedback analysis improves the system's fault tolerance in actuator failure scenarios, preventing flooding or energy loss caused by drain valve failure.

[0080] Optionally, after dynamically adjusting the drainage flow of the fuel cell drain valve according to the flooding risk level, the method further includes: determining the duration when the real-time entropy weight score is less than a first preset threshold and the anode gas pressure difference change rate and the stack voltage fluctuation amplitude are both less than the corresponding thresholds; and releasing the flooding mode when the duration is greater than the preset time threshold.

[0081] Specifically, after dynamically adjusting the drain flow rate of the fuel cell drain valve based on the flooding risk level, the controller further determines whether to release the flooding mode. Specifically, the controller continuously monitors the real-time entropy weight score, the rate of change of the anode gas pressure difference, and the stack voltage fluctuation amplitude. When the real-time entropy weight score is less than a first preset threshold (i.e., less than 0.3), indicating that the flooding risk has been significantly reduced, and the rate of change of the anode gas pressure difference is less than the corresponding rate of change threshold, which can be 1 kPa / s, indicating that the rate of change of the gas pressure difference over time is within a stable range, and the stack voltage fluctuation amplitude is also less than the corresponding threshold, which can be ±1%, indicating that the voltage fluctuation is within the normal operating range, a timer is started to determine the duration of this stable state. If the duration is greater than a preset time threshold, which can be 3 seconds, it is determined that the anode flooding has been effectively released, and the fuel cell system returns to normal operating mode. At this point, the flooding mode is released, and special control strategies for flooding, such as dynamic power allocation and enhanced drainage, are discontinued, allowing the fuel cell system to return to normal energy and water management processes.

[0082] In addition, when the flooding mode is released, it will switch to the transition mode to gradually restore the FCS power allocation weight to the normal level to avoid sudden changes in system parameters. The power allocation weight is smoothly adjusted in the transition mode:

[0083] ;

[0084] in, Indicates the lithium battery power allocation weight at the current time t in transition mode, ranging from 0 to 1, indicating the proportion of the load power borne by the lithium battery to the total power. Indicates the power distribution weight of the lithium battery in flooding mode, that is, the default weight of the lithium battery when the system is flooded. Indicates the lithium battery power distribution weight in normal operating mode, that is, the default weight when the system is running stably after flooding is resolved. Indicates the transition time constant in seconds or minutes, which is used to control the smooth switching speed from fault mode to normal mode. The larger it is, the slower the weight adjustment is and the smoother the system transition is. The duration of the transition process, i.e., the time elapsed from the moment flooding is released to the current time t, is represented by the transition mode. This prevents the risk of new flooding caused by sudden power changes when releasing flooding mode. Continuous model optimization through data accumulation allows adaptation to the individual characteristics of different fuel cell systems.

[0085] The technical solution of the embodiment of the present invention can comprehensively and accurately reflect the flooding risk status of the fuel cell anode by constructing a dynamic feature set and calculating a real-time entropy weight score, thereby improving the early warning capability and diagnostic accuracy of flooding failures. The output power, loading rate, and drainage flow are dynamically adjusted based on the real-time entropy weight score, realizing adaptive control of fuel cell water management and avoiding the lag and over-adjustment problems of traditional fixed threshold control strategies. It reduces dependence on manual experience and automatically identifies the flooding risk level through a data-driven entropy weight analysis method, reducing the probability of flooding failures due to subjective misjudgment, enhancing the reliability and durability of the fuel cell system, and effectively extending the life of the stack.

[0086] Example 2

[0087] Figure 2 This is a flow chart of a fuel cell flooding control method provided by the second embodiment of the present invention. This embodiment adds a specific process of identifying the fault type by real-time entropy weight scoring based on the above-mentioned first embodiment. Figure 2 As shown, the method includes:

[0088] S210: Obtain historical entropy weight scores, and calculate the score deviation between the real-time entropy weight scores and the historical entropy weight scores.

[0089] It can be known that during the operation of the fuel cell, in order to ensure system safety and diagnostic accuracy, the entropy weight model and sensor status need to be monitored in real time.

[0090] Specifically, the controller can obtain the entropy weighted scores of the last K historical moments. The historical scores represent the risk assessment results of the fuel cell under recent normal and potentially abnormal operating conditions. The controller will arithmetic average the K historical entropy weighted scores to obtain the average of the historical entropy weighted scores. It then calculates the absolute deviation between the real-time score and the historical average and presets a score deviation threshold, which can be 0.2.

[0091] S220: Obtain the historical distribution corresponding to each standard feature, and calculate the individual deviation corresponding to each standard feature based on the historical distribution.

[0092] Specifically, the operating status of the fuel cell is reflected by multiple standard features, including air pressure difference, air pressure difference change rate, and voltage fluctuation value. For each standard feature, the system constructs a probability density function based on the historical data in the sliding window. For example, for the air pressure difference feature, the sliding window contains N air pressure difference values ​​collected continuously over a period of time in the past, such as [10kPa, 10.2kPa, 9.8kPa, 10.1kPa…]. By performing statistical analysis on the data in the window, a probability density function that conforms to its distribution law, such as Gaussian distribution, can be constructed. The controller can then calculate the standardized distance between each current standard feature value and the corresponding historical distribution expectation as a separate deviation.

[0093] S230: When the score deviation is greater than a preset score deviation threshold and each individual deviation is within a preset normal range, it is determined that the entropy weight model is invalid.

[0094] Specifically, when the score changes suddenly but the features are normal, it indicates that the entropy weight model parameters are outdated and need to be updated. At this time, the model's self-learning mechanism is triggered. This mechanism will re-collect more representative fuel cell operation data over a period of time and use a specific algorithm to recalculate the weights of each standard feature in the entropy weight model, so that the model can more accurately assess the flooding risk of the fuel cell.

[0095] S240: When one of the individual deviations is greater than a preset high abnormality threshold, it is determined that the sensor is faulty.

[0096] Specifically, when a single feature fluctuates abnormally while other features are normal, it conforms to the sensor single-point failure mode, that is, the sensor may be subject to external interference, hardware damage, etc., which causes the data it collects to deviate, thereby affecting the individual deviation calculated based on the data. If the fuel cell system is equipped with redundant sensors, the system will immediately enable the redundant sensor to replace the faulty sensor to continue collecting data to ensure the normal operation of the system. At the same time, the system will mark the location and type of the faulty sensor in the fault diagnosis log, such as "air pressure difference sensor failure, located at the anode entrance", and trigger the corresponding alarm information to notify maintenance personnel to repair or replace it in time. In some vehicle application scenarios, the driver may be prompted with sensor fault information through the on-board display screen, such as "air pressure difference sensor abnormality, please repair as soon as possible."

[0097] S250: When the plurality of individual deviations are all greater than the preset abnormality threshold, a fuel cell system-level fault is determined.

[0098] Specifically, when multiple features are abnormal at the same time, it is consistent with systemic faults such as flooding and membrane electrode aging. When it is determined to be a system-level fault, the system will immediately enter safety mode and limit the power output of the fuel cell to prevent the fault from further deteriorating. At the same time, combined with the flooding risk level reflected by the real-time entropy weight score, the corresponding fault tolerance strategy is activated. For example, if the current flooding risk level is moderate flooding, the system may increase the drainage frequency of the drain valve to alleviate the flooding situation; if the risk level is severe flooding, in addition to increasing the drainage frequency, it may also reduce the operating current of the fuel cell and reduce the water production to maintain the stable operation of the system as much as possible until the fault is effectively detected and repaired.

[0099] Furthermore, fuel cells exhibit varying operating characteristics and fault sensitivity under different operating conditions. Therefore, fault diagnosis thresholds require dynamic adjustment based on operating conditions to improve diagnostic accuracy and reliability. This includes adjusting thresholds for high-load conditions and low-temperature startup. High-load threshold adjustment refers to the fact that fuel cell system reactions are more intense and parameter fluctuations are relatively large under high-load conditions. To prevent false faults due to normal parameter fluctuations, the score deviation threshold is lowered to 0.15. For example, under high-load conditions such as vehicle acceleration uphill, which require high fuel cell power output, if the score deviation exceeds 0.15, the system will proceed with more cautious troubleshooting, rather than initiating in-depth analysis as it would if the score deviation exceeds 0.2 under normal operating conditions. Low-temperature startup threshold adjustment refers to the fact that during the fuel cell startup phase, the low temperature environment affects the electrochemical reaction rate and material properties within the cell, resulting in differences in sensor data and system parameters compared to normal operation at room temperature. To accommodate this, the sensor deviation threshold is relaxed to 3.5 standard deviations. For example, in cold winter conditions, when a fuel cell starts, sensor data such as air pressure difference may fluctuate significantly due to changes in the physical properties of the gas caused by low temperatures. At this time, loosening the threshold can avoid misjudging sensor failures due to these normal low-temperature startup characteristics.

[0100] It's important to note that to prevent misjudgments due to transient interference, all fault determinations require that the fault condition be met for at least 200 ms. For example, if, at a certain moment, a single deviation in voltage fluctuation due to external electromagnetic interference momentarily exceeds the preset high anomaly threshold, but persists for only 50 ms, the system will not identify this as a sensor or system-level fault. Instead, it will wait to see if the anomaly persists. Only if the anomaly persists for 200 ms or longer will the system determine a fault based on the corresponding determination criteria.

[0101] The technical solution of the embodiment of the present invention can quickly detect abnormal fluctuations in scores by comparing real-time and historical entropy weight scores, providing a basis for fault warning. Individual deviation calculation analyzes the degree of data deviation from each standard feature dimension to accurately locate the source of the anomaly. Fault type determination is based on score deviation and individual deviation, scientifically distinguishing between entropy weight model failure, sensor failure, and system-level failure, facilitating targeted treatment. Dynamic threshold adaptation flexibly adjusts the judgment threshold according to operating conditions, reducing the misjudgment rate under different operating conditions. The fault confirmation mechanism effectively avoids misjudgments caused by transient interference by setting a duration and voting mechanism. This achieves comprehensive, accurate, and reliable fault diagnosis and treatment for the fuel cell system, ensuring stable system operation and improving its safety and reliability.

[0102] Example 3

[0103] Figure 3 This is a schematic diagram of the structure of a fuel cell flooding control device provided by the third embodiment of the present invention. The device can be implemented in software and / or hardware and can generally be integrated into the electronic device that executes the method. Figure 3 As shown, the device includes: a dynamic feature set construction module 310, which is used to obtain the feature vector of the fuel cell and input the feature vector into a dynamically updated sliding window one by one to form a dynamic feature set containing the features of the most recent N moments, where N is the preset storage capacity of the sliding window;

[0104] Scoring and level determination module 320, for calculating the real-time entropy weight score corresponding to the dynamic feature set, and determining the flooding risk level of the fuel cell anode according to the real-time entropy weight score;

[0105] An output power and loading rate adjustment module 330 is configured to dynamically adjust the output power and loading rate of the fuel cell according to the real-time entropy weight score when the real-time entropy weight score is greater than a first preset threshold;

[0106] The drainage flow regulating module 340 is used to dynamically regulate the drainage flow of the fuel cell drainage valve according to the flooding risk level.

[0107] Optionally, the dynamic feature set construction module 310 is specifically used to: collect the inlet and outlet air pressures and the stack voltage of the fuel cell anode; calculate the inlet and outlet air pressure difference of the anode based on the inlet and outlet air pressures, and determine the air pressure difference change rate of the anode based on the inlet and outlet air pressure difference; determine the voltage fluctuation value based on the stack voltage; and form a feature vector based on the inlet and outlet air pressure difference, the air pressure difference change rate and the voltage fluctuation value.

[0108] Optionally, the scoring and grade determination module 320 specifically includes: a real-time entropy weight scoring calculation unit, which is used to: perform range normalization on each window feature in the dynamic feature set to generate a standard feature; take each standard feature as the target feature in turn, calculate the feature normalization ratio based on the target feature, and calculate the information entropy according to the feature normalization ratio; determine the feature weight of the target feature based on the information entropy calculation; and perform weighted summation of each target feature and its corresponding feature weight through an entropy weight model to determine the real-time entropy weight score.

[0109] Optionally, the device also includes: a fault determination module, which is used to: obtain historical entropy weight scores, calculate the score deviation between the real-time entropy weight scores and the historical entropy weight scores; obtain the historical distribution corresponding to each standard feature, and calculate the individual deviation corresponding to each standard feature according to the historical distribution; when the score deviation is greater than the preset score deviation threshold and each individual deviation is within the preset normal range, it is determined that the entropy weight model has failed; when one of the individual deviations is greater than the preset high abnormality threshold, it is determined that the sensor has failed; when multiple individual deviations are greater than the preset abnormality threshold, it is determined that the fuel cell system level has failed.

[0110] Optionally, the scoring and level determination module 320 specifically includes: a risk level determination unit, used to: when the real-time entropy weight score is less than or equal to a first preset threshold, determine that the flood risk level is normal; when the real-time entropy weight score is greater than the first preset threshold and less than or equal to the second preset threshold, determine that the flood risk level is mild flooding; when the real-time entropy weight score is greater than the second preset threshold and less than or equal to the third preset threshold, determine that the flood risk level is moderate flooding; when the real-time entropy weight score is greater than the third preset threshold, determine that the flood risk level is severe flooding.

[0111] Optionally, the output power and loading rate adjustment module 330 is specifically used to: obtain the basic weight and weight gain coefficient of the lithium battery, calculate the product of the real-time entropy weight score and the weight gain coefficient, and add the product to the basic weight to obtain the priority weight of the lithium battery; determine the target weight of the fuel cell according to the priority weight of the lithium battery; obtain the total load power, and distribute the output power of the fuel cell according to the product of the target weight and the total load power; obtain the anode gas pressure difference change rate, and calculate the loading rate of the fuel cell according to the anode gas pressure difference change rate and the real-time entropy weight score; wherein the loading rate is calculated using the following formula:

[0112] ;

[0113] in, is the loading rate, is the preset maximum loading rate, is the preset safety threshold pressure difference change rate, Score the real-time entropy weight, is the anode gas pressure difference change rate, and t is the current moment.

[0114] Optionally, the drainage flow regulation module 340 is specifically used to: determine the basic drainage coefficient corresponding to the flooding risk level, obtain the target change of the real-time status parameters, wherein the real-time status parameters include the anode gas pressure difference change rate and the stack voltage fluctuation amplitude change rate; obtain fuzzy inference rules, wherein the fuzzy inference rules include the drainage correction amount corresponding to each change amount; correct the basic drainage coefficient by the drainage correction amount to generate a target drainage coefficient, and determine the drainage flow of the fuel cell drain valve according to the target drainage coefficient.

[0115] Optionally, the device also includes: a flooding risk release module, which is used to: after dynamically adjusting the drainage flow of the fuel cell drain valve according to the flooding risk level, determine the duration when the real-time entropy weight score is less than a first preset threshold and the anode gas pressure difference change rate and the stack voltage fluctuation amplitude are both less than the corresponding threshold; when the duration is greater than the preset time threshold, release the flooding mode.

[0116] The technical solution of the embodiment of the present invention can comprehensively and accurately reflect the flooding risk status of the fuel cell anode by constructing a dynamic feature set and calculating a real-time entropy weight score, thereby improving the early warning capability and diagnostic accuracy of flooding failures. The loading rate and drainage flow are dynamically adjusted based on the real-time entropy weight score, realizing adaptive control of fuel cell water management, avoiding the lag and over-adjustment problems of traditional fixed threshold control strategies, and improving the system response speed and stability. It reduces dependence on manual experience, automatically identifies the flooding risk level through a data-driven entropy weight analysis method, reduces the probability of flooding failures caused by subjective misjudgment, enhances the reliability and durability of the fuel cell system, and effectively extends the life of the stack.

[0117] A fuel cell flooding control device provided by an embodiment of the present invention can execute a fuel cell flooding control method provided by any embodiment of the present invention, and has corresponding functional modules and beneficial effects of the execution method.

[0118] Example 4

[0119] Figure 4 A schematic diagram of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices (such as helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0120] like Figure 4 As shown, electronic device 10 includes at least one processor 11 and memory, such as read-only memory (ROM) 12 and random access memory (RAM) 13, communicatively connected to at least one processor 11. The memory stores computer programs executable by the at least one processor. Processor 11 can perform various appropriate actions and processes based on the computer programs stored in ROM 12 or loaded from storage unit 18 into RAM 13. RAM 13 can also store various programs and data required for the operation of electronic device 10. Processor 11, ROM 12, and RAM 13 are interconnected via bus 14. An input / output (I / O) interface 15 is also connected to bus 14.

[0121] Multiple components in the electronic device 10 are connected to the I / O interface 15, including an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0122] Processor 11 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 executes the various methods and processes described above, such as a fuel cell flooding control method. Specifically, the method includes: obtaining a fuel cell feature vector, sequentially inputting the feature vector into a dynamically updated sliding window to form a dynamic feature set containing the features of the window at the most recent N moments, where N is the preset storage capacity of the sliding window; calculating a real-time entropy weight score corresponding to the dynamic feature set, and determining the flooding risk level of the fuel cell anode based on the real-time entropy weight score; dynamically adjusting the fuel cell output power and loading rate based on the real-time entropy weight score when the real-time entropy weight score is greater than a first preset threshold; and dynamically adjusting the drain flow rate of the fuel cell drain valve based on the flooding risk level.

[0123] In some embodiments, a fuel cell flooding control method can be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the fuel cell flooding control method described above can be performed. Alternatively, in other embodiments, processor 11 can be configured to execute a fuel cell flooding control method in any other suitable manner (e.g., via firmware).

[0124] Various embodiments of the systems and techniques described above can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0125] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer program may be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0126] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, device, or apparatus. A computer-readable storage medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or apparatus, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media may include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0127] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device that has: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0128] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0129] A computing system may include clients and servers. The clients and servers are generally remote from each other and typically interact via a communication network. This client-server relationship arises through computer programs running on the respective computers, creating a client-server relationship. The server may be a cloud server, also known as a cloud computing server or cloud host. This server is a hosting product within the cloud computing service ecosystem that addresses the management difficulties and limited scalability of traditional physical hosts and VPS servers.

[0130] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in the present invention can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved. This is not limited herein.

[0131] The above specific embodiments do not limit the scope of protection of the present invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A fuel cell flooding control method, characterized in that: include: Obtaining a feature vector of the fuel cell, and inputting the feature vector into a dynamically updated sliding window one by one to form a dynamic feature set containing features of the most recent N time windows, where N is a preset storage capacity of the sliding window; Calculating a real-time entropy weight score corresponding to the dynamic feature set, and determining a flooding risk level of the fuel cell anode according to the real-time entropy weight score; When the real-time entropy weight score is greater than a first preset threshold, dynamically adjusting the output power and loading rate of the fuel cell according to the real-time entropy weight score; dynamically adjusting the drainage flow of the fuel cell drain valve according to the flooding risk level; The step of obtaining a characteristic vector of a fuel cell includes: Collect the inlet and outlet gas pressures and stack voltage of the fuel cell anode; Calculating the inlet and outlet air pressure difference of the anode according to the inlet and outlet air pressures, and determining the air pressure difference change rate of the anode according to the inlet and outlet air pressure difference; determining a voltage fluctuation value according to the stack voltage; A characteristic vector is formed according to the inlet and outlet pressure difference, the pressure difference change rate and the voltage fluctuation value; The step of calculating the real-time entropy weight score corresponding to the dynamic feature set includes: performing range normalization on each window feature in the dynamic feature set to generate a standard feature; Taking each standard feature as a target feature in turn, calculating a feature normalization ratio based on the target feature, and calculating information entropy according to the feature normalization ratio; Determine the feature weight of the target feature according to the information entropy calculation; The target features and their corresponding feature weights are weighted and summed using an entropy weight model to determine a real-time entropy weight score; wherein the output power and loading rate of the fuel cell are dynamically adjusted according to the real-time entropy weight score, including: Obtaining a basic weight and a weight gain coefficient of the lithium battery, calculating the product of the real-time entropy weight score and the weight gain coefficient, and adding the product to the basic weight to obtain a priority weight of the lithium battery; determining a target weight for the fuel cell based on the lithium battery priority weight; Obtaining total load power, and allocating the output power of the fuel cell according to the product of the target weight and the total load power; Obtaining a change rate of the anode gas pressure difference, and calculating a loading rate of the fuel cell according to the change rate of the anode gas pressure difference and the real-time entropy weight score; The loading rate is calculated using the following formula: ； in, is the loading rate, is the preset maximum loading rate, is the preset safety threshold pressure difference change rate, Score the real-time entropy weight, is the anode gas pressure difference change rate, t is the current moment; The dynamically adjusting the drainage flow of the fuel cell drain valve according to the flooding risk level includes: Determine a basic drainage coefficient corresponding to the flooding risk level, and obtain a target change in real-time state parameters, wherein the real-time state parameters include a change rate of the anode gas pressure difference and a change rate of the stack voltage fluctuation amplitude; Obtaining a fuzzy inference rule, wherein the fuzzy inference rule includes a drainage correction amount corresponding to each change amount; The basic drainage coefficient is corrected by the drainage correction amount to generate a target drainage coefficient, and the drainage flow rate of the fuel cell drainage valve is determined according to the target drainage coefficient.

2. The method according to claim 1, characterized in that The method further comprises: Obtaining historical entropy weight scores, and calculating a score deviation between the real-time entropy weight score and the historical entropy weight score; Obtaining a historical distribution corresponding to each standard feature, and calculating a separate deviation corresponding to each standard feature based on the historical distribution; When the score deviation is greater than a preset score deviation threshold and each individual deviation is within a preset normal range, it is determined that the entropy weight model is invalid; When one of the individual deviations is greater than the preset high abnormality threshold, the sensor is determined to be faulty; When multiple individual deviations are greater than a preset abnormality threshold, a fuel cell system-level fault is determined.

3. The method according to claim 1, characterized in that Determining the flooding risk level of the fuel cell anode according to the real-time entropy weight score includes: When the real-time entropy weight score is less than or equal to a first preset threshold, determining that the flood risk level is normal; When the real-time entropy weight score is greater than a first preset threshold and less than or equal to a second preset threshold, the flood risk level is determined to be mild flooding; When the real-time entropy weight score is greater than the second preset threshold and less than or equal to the third preset threshold, the flood risk level is determined to be moderate flooding; When the real-time entropy weight score is greater than a third preset threshold, the flood risk level is determined to be severe flooding.

4. The method according to claim 1, wherein After dynamically adjusting the drainage flow of the fuel cell drainage valve according to the flooding risk level, the method further includes: When the real-time entropy weight score is less than a first preset threshold and the anode gas pressure difference change rate and the stack voltage fluctuation amplitude are both less than corresponding thresholds, determining a duration; When the duration is greater than the preset time threshold, the flooding mode is released.

5. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively coupled to the at least one processor; The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 4.

6. A computer storage medium, characterized in that The computer storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the method according to any one of claims 1 to 4 when executed.

Images