Fuel cell state estimation method, state evaluation device and storage medium

Abstract

The present invention provides a fuel cell state estimation method, state assessment device, and storage medium, including the following steps: operating mode classification; establishing a comprehensive evaluation index system; operating the fuel cell at different stages, obtaining the values ​​of the accuracy index and stability index in the observer comprehensive evaluation index system, and normalizing them; determining the accuracy score and stability score of each sub-index under each operating mode; using a combined weighted multiplication synthesis method to determine the accuracy sub-index score and stability sub-index score of different observers under each operating mode; determining the comprehensive score of different observers used in different stages; and establishing a fusion observer control strategy based on the comprehensive score of each observer at different stages. The evaluation results of the present invention are objective and reliable, and can provide direction for improving fuel cell state estimation.

Description

Technical Field

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

[0002] Fuel cells need to frequently cope with complex operating conditions such as dynamic vehicle loading, start-stop operations, continuous low-load or idling operation, and high power output. To ensure stable operation and improve efficiency of fuel cells, real-time monitoring of the fuel cell status is required.

[0003] Currently, fuel cell system observers include Luenberger observer, Kalman filter observer (KF), sliding mode observer (SMO), etc. These observers play an important role in estimating the state of PEMFC system.

[0004] While a single observer can provide a certain level of state estimation capability under certain conditions, its limitations gradually become apparent in complex practical applications. When the fuel cell's workload, temperature, or fuel quality change, a single observer may not be able to accurately capture these changes, leading to deviations in state estimation. Secondly, a single observer lacks redundancy. Once a failure or performance degradation occurs, it will directly affect the operation of the entire system, reducing its reliability and stability. Furthermore, due to the lack of diverse sources of observation data, a single observer has limited state estimation and fault detection capabilities, making it difficult to detect and handle abnormal situations in a timely manner, further increasing the risk of system operation.

[0005] In contrast, multi-observer fusion strategies demonstrate significant advantages in fuel cell state estimation. By integrating observations from multiple observers, multi-observer fusion significantly improves state estimation accuracy. Different observers are based on distinct assumptions or modeling approaches, and their complementary effects help reduce errors caused by bias in a single observer, thereby providing more accurate state estimates. Summary of the Invention

[0006] In response to the shortcomings of the existing technology, the present invention provides a fuel cell state estimation method, a fuel cell state evaluation device and a storage medium, which can establish a comprehensive score based on the observer accuracy index and stability index, and is suitable for fuel cell state estimation under different operating modes.

[0007] The present invention achieves the above technical objectives through the following technical means.

[0008] A fuel cell state estimation method comprises the following steps:

[0009] Working mode division: According to the change of working current, the working mode is divided into starting stage, steady state stage and variable load stage;

[0010] Establish a comprehensive evaluation index system, which includes an observer accuracy comprehensive evaluation matrix G and a stability comprehensive evaluation matrix D; the observer accuracy comprehensive evaluation matrix G includes the weights of accuracy indicators at different working stages, and the observer stability comprehensive evaluation matrix D includes the weights of stability indicators at different working stages;

[0011] The fuel cell is operated at different stages to obtain and normalize the values ​​of the accuracy index and the stability index in the observer comprehensive evaluation index system; the values ​​of the accuracy index include the root mean square error value, the mean absolute percentage error value, the mean absolute percentage error, and the coefficient of determination; the values ​​of the stability index include the maximum convergence time, the average convergence time, the maximum jitter amplitude, and the average jitter amplitude; the normalized values ​​of the accuracy sub-indicators are used to form an accuracy index normalization matrix J, and the normalized values ​​of the stability sub-indicators are used to form a stability index normalization matrix H;

[0012] The sub-indicator observer accuracy comprehensive evaluation matrix G, stability comprehensive evaluation matrix D and the obtained accuracy normalized matrix The matrix homogenized with the stability sub-index Conduct a comprehensive evaluation to determine the accuracy score of each sub-indicator under each working mode and stability score ;

[0013] The accuracy sub-index score and stability sub-index score of the SMO observer under each working mode are determined by the combined weighting-multiplication synthesis method; the accuracy sub-index score and stability sub-index score of the UKF observer under each working mode are determined by the combined weighting-multiplication synthesis method;

[0014] Determine the comprehensive scoring using SMO observer and the comprehensive scoring using UKF observer at different stages;

[0015] The control strategy of the fusion observer is established based on the comprehensive score of each observer at different stages.

[0016] Furthermore, the comprehensive evaluation matrix G of the observer accuracy is expressed as: , express The weight of the u-th accuracy indicator of the stage; , u=1 represents the root mean square error; u=2 represents the mean absolute percentage error; u=3 represents the maximum error; u=4 represents the coefficient of determination; , Indicates the startup phase, represents the steady-state phase, Indicates the load change stage;

[0017] The comprehensive evaluation matrix D of observer stability is expressed as: , express The weight of the zth stability indicator of the stage; , z=1 represents the maximum convergence time; z=2 represents the average convergence time; z=3 represents the maximum jitter amplitude; z=4 represents the average jitter amplitude.

[0018] Furthermore, the fuel cell is operated at different stages to obtain the value of the accuracy index in the observer comprehensive evaluation index system. , , where: A1 is the root mean square error value, A2 is the mean absolute percentage error value, A3 is the mean absolute percentage error; A4 is the determination coefficient;

[0019] ;

[0020] ;

[0021] ;

[0022] ;

[0023] Where: n represents the number of samples collected, m represents the position of each observation in the data set, Indicates the actual value, represents the observation value using the unscented Kalman filter observer, represents the predicted value of the observation using the unscented Kalman filter observer, Indicates the average value of the actual value. k indicates the number of times the data is collected; t k Indicates the time of the kth data collection; t k+1 Indicates the time of the data collected for the k+1th time;

[0024] Run the fuel cell at different stages to obtain the value of the stability index in the observer comprehensive evaluation index system , Where: X1 is the maximum convergence time; X2 is the average convergence time; X3 is the maximum jitter amplitude; X4 is the average jitter amplitude;

[0025] ;

[0026] ;

[0027] ;

[0028] ;

[0029] in, represents the time taken for data convergence in the dth data collection, and e represents the total number of data collections; is the maximum value of the observation value using the unscented Kalman filter observer, is the minimum value of the observation value using the unscented Kalman filter observer;

[0030] The values ​​of sub-indicators at different stages and Perform homogenization to obtain the homogenized values ​​of the accuracy sub-indicators to form the accuracy index homogenization matrix The stability index normalization matrix is ​​formed by the normalized values ​​of the stability sub-indicators ,in:

[0031] ;

[0032] ;

[0033] Where: express The value of the u-th accuracy index after normalization using the observer in the stage; express The normalized value of the zth stability index after using the observer in the stage.

[0034] Further, determine the accuracy score of each sub-indicator under each working mode , ; Accuracy score of each sub-indicator in each working mode Expressed as , Represented as the vth column of the matrix;

[0035] Determine the stability score of each sub-indicator under each working mode , ; Accuracy score of each sub-indicator in each working mode Expressed as , Represented as the w-th column of the matrix.

[0036] Furthermore, a combined weighted-multiplication synthesis method is used to determine the accuracy sub-index score and stability sub-index score of the SMO observer under each working mode; a combined weighted-multiplication synthesis method is used to determine the accuracy sub-index score and stability sub-index score of the UKF observer under each working mode. Specifically:

[0037] The accuracy score of each sub-indicator under each working mode obtained by the SMO observer is recorded as ; The accuracy index normalization matrix under each working mode is obtained by SMO observer and recorded as ; The stability score of each sub-index under each working mode obtained by the SMO observer is recorded as ; The stability index normalized matrix under each working mode is obtained by SMO observer and recorded as ; The accuracy score matrix of each sub-indicator under each working mode obtained by the SMO observer and the observer accuracy normalized value matrix The combined weighted-multiplication synthesis method is used to calculate and obtain The accuracy score of the u-th sub-indicator in the stage is recorded as ; The stability score matrix of each sub-index under each working mode obtained by the SMO observer and the observer stability normalization matrix The combined weighted-multiplication synthesis method is used to calculate and obtain The stability score of the zth sub-index in the stage is recorded as ;

[0038] The accuracy score of each sub-indicator under each working mode obtained by the UKF observer is recorded as , the accuracy index normalized matrix under each working mode is obtained through the UKF observer and recorded as ; The stability index normalized matrix under each working mode is obtained by UKF observer and recorded as ; The accuracy score matrix of each sub-indicator under each working mode obtained by the UKF observer and the observer accuracy normalization matrix The combined weighted-multiplication synthesis method is used to calculate and obtain the The accuracy score of the u-th sub-indicator in the stage is recorded as ; The stability score matrix of each sub-index under each working mode obtained by the UKF observer and the observer stability normalization matrix The combined weighted-multiplication synthesis method is used to calculate and obtain the The stability score of the zth sub-index in the stage is recorded as .

[0039] Furthermore, the comprehensive scoring using the SMO observer and the comprehensive scoring using the UKF observer were determined at different stages, specifically:

[0040] According to the SMO observer The accuracy score of the u-th sub-indicator at stage and the stability score of the zth sub-index Sure SMO observer comprehensive score of the stage:

[0041] ;

[0042] According to the UKF observer The accuracy score of the u-th sub-indicator at stage and the stability score of the zth sub-index Sure UKF observer comprehensive score for the stage:

[0043] ;

[0044] Where:

[0045] for The stage was scored comprehensively using the SMO observer; for The SMO observer was used to comprehensively score the stage.

[0046] Furthermore, the control strategy of the fusion observer is established based on the comprehensive score of each observer at different stages, specifically:

[0047] exist In the stage, the comprehensive score of the state estimation using the SMO observer is compared Comprehensive score with state estimation using UKF observer ,

[0048] like ,and , then in In this stage, the SMO observer is used as the observer for fuel cell state estimation to perform performance evaluation;

[0049] like ,and , then use the fusion observation strategy;

[0050] like ,and , then in In this stage, the UKF observer is used as the observer for fuel cell state estimation to perform performance evaluation;

[0051] like ,and , then the fusion observation strategy is used; a is the set value.

[0052] Furthermore, the fusion observation strategy is specifically as follows:

[0053] The accuracy score and stability score using the SMO observer and UKF observer are calculated respectively, and the calculation formulas are as follows:

[0054] ;

[0055] ;

[0056] Where: For The accuracy score of the stage was scored using the SMO observer; For The stability score of the stage using the SMO observer; For The accuracy score of the UKF observer was used in the stage; For The stability score of the UKF observer is used in the stage;

[0057] when > ,and > , then the SMO observer is used for performance evaluation;

[0058] when < ,and < , then the UKF observer is used for performance evaluation;

[0059] when > ,and < , the SMO observer is used for accuracy performance evaluation, and the UKF observer is used for stability performance evaluation;

[0060] when < ,and > , the UKF observer is used for accuracy performance evaluation, and the SMO observer is used for stability performance evaluation.

[0061] A fuel cell state evaluation device comprises a processor and a memory, wherein the memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the steps in the fuel cell state estimation method are executed.

[0062] A storage medium stores a computer program, which runs the steps of the fuel cell state estimation method when executed by a processor.

[0063] The beneficial effects of the present invention are:

[0064] 1. The fuel cell state estimation method described in this invention aims to comprehensively consider the accuracy and stability of fuel cell state assessments during different operating phases. By constructing a comprehensive evaluation matrix applicable to each operating mode, this method can provide more comprehensive and accurate state monitoring and prediction, thereby ensuring the efficient, safe, and reliable operation of the fuel cell system.

[0065] 2. The fuel cell state estimation method of the present invention can obtain scores for evaluation indicators related to fuel cell hydrogen supply system state estimation, and can provide a reference direction for improving fuel cell state estimation.

[0066] 3. The fuel cell state estimation method described in this invention enables researchers and developers of fuel cell state estimation to obtain optimal observer evaluation results under different operating stages, thereby guiding the improved design of fuel cell hydrogen supply system state estimation components. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. The drawings described below are some embodiments of the present invention. For ordinary technicians in this field, it is obvious that other drawings can be obtained based on these drawings without paying any creative work.

[0068] Figure 1 This is a schematic diagram of the fuel cell operating mode division according to the present invention.

[0069] Figure 2 This is a flow chart of the fusion observer control strategy described in the present invention. DETAILED DESCRIPTION

[0070] The following describes embodiments of the present invention in detail, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and are not to be construed as limiting the present invention.

[0071] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "axial", "radial", "vertical", "horizontal", "inside", "outside" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0072] In the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," "connect," "fixed," etc. should be understood broadly. For example, they may refer to fixed, detachable, or integral connections; mechanical or electrical connections; direct or indirect connections through an intermediary; or internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0073] The fuel cell state estimation method of the present invention comprises the following steps:

[0074] S01: Working mode division, according to the change of working current, the working mode is divided into starting stage, steady state stage and variable load stage. Figure 1 As shown, the working modes are specifically divided into:

[0075] According to different loads, the working current of the fuel cell can be divided into n stages, where the current of stage i is set to I i , i∈(1,n);

[0076] Startup phase: load current I gradually increases from 0 to When the fuel cell is running, the load current I will go through a period of acceleration until it reaches a stable working state. During this stage, the current load changes greatly and the circuit performance may not be stable enough.

[0077] Steady state: for different loads, at least Current in time remains unchanged, which is called the i-th steady-state stage. Current in time remains unchanged, which is called the i+1 steady-state stage.

[0078] Load-variable phase: The process where the current changes from the i-th steady-state phase to the i+1-th steady-state phase. The change in load current between these two stable phases is defined as the load-variable phase, allowing for flexible adaptation to various load conditions.

[0079] S02: Establish a comprehensive evaluation index system, including constructing the observer accuracy comprehensive evaluation matrix G and the observer stability comprehensive evaluation matrix D;

[0080] The observer accuracy comprehensive evaluation matrix G includes the weights of the accuracy index at different working stages, and the observer stability comprehensive evaluation matrix D includes the weights of the stability index at different working stages;

[0081] The accuracy indicators include root mean square error RMSE, mean absolute percentage error MAPE, maximum error MaxError, determination coefficient R-squared, Indicates the weight of the accuracy index, u is taken as 1, 2, 3, 4; represent The weight of the root mean square error RMSE of the stage; represent The weight of the mean absolute percentage error (MAPE) of the stage; represent The weight of the maximum error Max Error in the stage; represent The weight of the stage determination coefficient R-squared. In order to reflect the accuracy and stability indicators of different stages, the superscripts 1, 2 and 3 are used to represent the three stages, where 1 represents the startup stage, 2 represents the steady-state stage, and 3 represents the variable load stage. represents the weight of the root mean square error in the startup phase; express The weight of the u-th accuracy metric at stage u.

[0082] The comprehensive evaluation matrix G of the observer accuracy can be expressed as:

[0083] ;

[0084] The stability indicators include the maximum convergence time MCT, the average convergence time ACT, the maximum jitter amplitude MJA, and the average jitter amplitude AJA. represents the weight of the stability index, z is taken as 1, 2, 3, 4; represent The weight of the maximum convergence time MCT of the stage; represent The weight of the average convergence time ACT of the stage; represent The weight of the maximum jitter amplitude MJA of the stage; represent The weight of the average jitter amplitude AJA of the stage. In order to reflect the accuracy and stability indicators of different stages, the superscripts 1, 2 and 3 are used to represent the three stages, where 1 represents the startup stage, 2 represents the steady-state stage, and 3 represents the variable load stage. The weight of the maximum convergence time MCT in the startup phase; express The weight of the zth stability indicator at stage.

[0085] The comprehensive evaluation matrix D of observer stability can be expressed as:

[0086] ;

[0087] The specific weights in the embodiment are determined as follows:

[0088] In the indicator weights of accuracy:

[0089] The startup phase is 0.4, is 0.3, is 0.2, is 0.1.

[0090] The root mean square error (RMSE) during the startup phase is weighted at 0.4. During the startup phase, accuracy assessment typically focuses on minimizing the error between the observer and the fuel cell system. RMSE is a commonly used evaluation metric that indicates the degree of discrepancy between the observer output and the actual value. RMSE is given a higher weight because it penalizes the sum of squared results across the entire dataset, making it crucial for detecting large errors during the startup phase.

[0091] The mean absolute percentage error (MAPE) during the startup phase is weighted at 0.3. MAPE is a relative error metric that reflects the relative error between the observer output and the actual value. During the startup phase, assessing the percentage error helps understand the observer's accuracy, especially when the percentage error is involved. The MAPE is given a moderate weight of 0.3 to balance the emphasis on the percentage error with the performance of the observer during the startup phase.

[0092] The maximum error during the startup phase is assigned a weight of 0.2. Max Error reflects the maximum error that the observer can experience during startup. During this phase, the system's dynamic response may result in transient errors that do not necessarily require highly accurate measurements. Although Max Error may not significantly impact overall performance, assigning it a weight of 0.2 helps assess the accuracy and robustness of the observer.

[0093] The coefficient of determination during the startup phase is weighted 0.1. R-squared measures the proportion of the observer output variation that can be explained. During the startup phase, the goodness of fit of the observer can help determine whether it accurately reflects changes in the fuel cell system's estimated values. Although R-squared may be relatively low during the startup phase, a small weight of 0.1 is assigned to R-squared to comprehensively account for the impact of goodness of fit on observer performance.

[0094] The weight assignment described in the startup observer performance evaluation is designed to balance factors such as accuracy, relative error, maximum error, and goodness of fit. By assigning appropriate weights, these factors can be comprehensively considered to determine whether the observer can provide accurate and robust current estimation during the startup phase.

[0095] The steady-state phase is 0.3, is 0.4, is 0.2, is 0.1.

[0096] The root mean square error (RMSE) in the steady-state phase is weighted at 0.3. RMSE is a useful metric for measuring the error between estimates and actual values. During steady-state operation, it is desirable for the observer to ensure the accuracy of the estimated values. A lower RMSE value indicates that the observer is accurately tracking changes in the fuel cell system. It is desirable for the observer to provide accurate estimates to ensure stable system operation. Therefore, a higher weight is assigned to RMSE to emphasize the importance of accuracy.

[0097] The weight of the mean absolute percentage error in the steady-state phase is 0.4. During steady-state operation, current values ​​may fluctuate to some extent, and the MAPE effectively measures the percentage error of the observer output relative to the actual value. A lower MAPE value indicates that the observer is able to provide more accurate estimates in the presence of large current fluctuations. By assigning a higher weight to the MAPE in the weighting distribution, the importance of the observer's accuracy during steady-state operation is emphasized.

[0098] The maximum error in the steady-state phase is weighted at 0.2. Monitoring the Max Error ensures that the observer provides adequate safety protection during steady-state operation, promptly detecting potential errors or anomalies. While RMSE and MAPE provide a measure of overall error, they ignore anomalies at individual observation points. Assigning a moderate weight to the Max Error helps ensure the observer's sensitivity to anomalies, thereby enhancing the safety of steady-state operation.

[0099] The coefficient of determination for the steady-state phase is weighted at 0.1. R-squared is used to assess the degree of correlation between the observer output and the actual current. A high R-squared value indicates that the observer is able to better estimate the system state, enhancing the observer's accuracy and reliability. Although R-squared may be relatively unimportant during steady-state operation, by assigning a small amount of weight to it, we can still consider the observer's goodness of fit as part of the evaluation.

[0100] The rationale for weighting observer performance during steady-state operation is to balance multiple metrics, including accuracy, safety, and goodness of fit. This weighting comprehensively considers various factors to ensure the observer provides accurate and safe predictions during steady-state operation while maintaining relevance to the fuel cell system state.

[0101] The load change stage is 0.2, is 0.4, is 0.3, is 0.1.

[0102] The root mean square error (RMSE) during the load-variable phase is weighted at 0.2. During this phase, changes in the current load can significantly impact the observer's accuracy. Under these conditions, the observer must adapt to these changes and estimate the fuel cell system's state as accurately as possible. RMSE, a commonly used error metric, indicates the degree of discrepancy between the observer output and the actual value. Giving the RMSE a moderate weight of 0.2 reflects the emphasis on the observer's adaptability and accuracy during the load-variable phase.

[0103] The mean absolute percentage error (MAPE) is weighted 0.4 during the load variation phase. During this phase, relative error metrics such as MAPE are crucial for evaluating observer performance. Current load changes can cause the relative error between the observer output and the actual value to change. Therefore, using MAPE as an evaluation metric provides a better understanding of the observer's relative accuracy under load variations. Giving MAPE a higher weight of 0.4 reflects the emphasis placed on relative error assessments of load variations.

[0104] The maximum error during the load variation phase is weighted at 0.3. During this phase, monitoring the observer's Max Error provides timely detection of anomalies and safety protection. Max Error reflects the observer's potential Max Error value. When system loads vary, Max Error monitoring can help detect potential errors or anomalies. Assigning a moderate weight of 0.3 to Max Error emphasizes the importance of observer accuracy and safety.

[0105] The coefficient of determination (R-squared) is weighted 0.1 during the load-varying phase. R-squared measures the correlation between the observer output and the actual current. By evaluating the relationship between the observer output and the actual value during the load-varying phase, we can understand the reliability and consistency of the observer. Although the R-squared weight is lower at 0.1, it still provides a comprehensive assessment of the observer's performance.

[0106] Based on the above, the comprehensive evaluation matrix of observer accuracy G= .

[0107] In the indicator weight of stability:

[0108] Weight of stability during the startup phase 0.35, 0.25, 0.2, is 0.2.

[0109] The maximum convergence time during the startup phase is weighted at 0.35. MCT is the maximum time required for the observer to reach steady state from startup. A higher weight (0.35) is given to MCT because reaching steady state quickly during startup is crucial for system reliability and responsiveness. A shorter MCT indicates that the observer can quickly adapt to system changes and provide accurate outputs.

[0110] The average convergence time during the startup phase is weighted 0.25. ACT is the average time it takes for the observer to reach a stable state over multiple startups. The ACT is weighted 0.25 to reflect the stability of the average. This metric contributes to the overall performance of the observer during the startup phase. A shorter ACT indicates that the observer exhibits consistent stability across different startup conditions.

[0111] The maximum jitter amplitude during the startup phase has a weight of 0.2. Jitter amplitude refers to the maximum oscillation amplitude of the observer output during the startup phase. Considering the observer's interference rejection capability, the MJA metric is of particular importance. A weight of 0.2 is assigned to MJA to assess the observer's ability to suppress interference signals. A smaller MJA indicates better interference rejection and stable output.

[0112] The average jitter amplitude during the startup phase is weighted 0.2. AJA is the average jitter amplitude of the observer over multiple startups. A weight of 0.2 is assigned to AJA to ensure that the observer can suppress jitter under different startup conditions. A smaller AJA indicates better stability of the observer and can reduce fluctuations in the system output.

[0113] In summary, such a distribution can comprehensively consider the stability, response speed and anti-interference ability of the observer, ensuring reliable output during the startup phase of the fuel cell.

[0114] Weight of stability in the steady-state phase 0.2, 0.3, 0.25, is 0.25.

[0115] The maximum convergence time in the steady-state phase has a weight of 0.2. MCT specifies the maximum time required for the observer to reach a stable state from any initial state during steady-state operation. The low weight of 0.2 for MCT is given because during steady-state operation, the focus is on ensuring that the observer stabilizes quickly, and a smaller MCT provides faster response and greater reliability.

[0116] The average convergence time in the steady-state phase is weighted 0.3. ACT is the average time it takes the observer to reach a stable state over multiple steady-state starts. During steady-state operation, the ACT is weighted higher by 0.3 to ensure consistent stability across different operating conditions. A shorter ACT indicates that the observer can quickly adapt to system changes and provide stable outputs.

[0117] The weight of the maximum chattering amplitude during the steady-state phase is set to 0.25. MJA indicates the maximum oscillation amplitude that occurs during the observer's steady-state operation. Giving MJA a moderate weight of 0.25 reflects the importance of the observer's anti-chattering performance. A smaller MJA indicates that the observer has better ability to suppress chattering and interference.

[0118] The weight of the average jitter amplitude during the steady-state phase is set to 0.25. AJA is the average jitter amplitude of the observer over multiple steady-state startups. By assigning AJA a weight of 0.25, the stability level of the observer under different operating conditions can be evaluated. A smaller AJA indicates that the observer is able to provide stable output during the steady-state phase.

[0119] Weight of stability during load change phase 0.15, 0.25, 0.3, is 0.3.

[0120] The maximum convergence time in the load-varying phase has a weight of 0.15. During the load-varying phase, the MCT has a weight of 0.15, which means that the observer must maintain good stability even under extreme load variations. The MCT metric focuses more on performance under abnormal conditions.

[0121] The average convergence time during the load variation phase is weighted 0.25. During this phase, the observer's stability and rapid response become crucial. The ACT is given a higher weight of 0.25 because during this phase, the observer must be able to quickly adapt to load changes and converge to a stable state to provide accurate output. The ACT weight of 0.25 reflects the importance of both observer stability and responsiveness during load variations.

[0122] The maximum jitter amplitude during the load-varying phase is weighted at 0.3, while the average jitter amplitude during the load-varying phase is weighted at 0.3. Both MJA and AJA have weights of 0.3, indicating that the observer's anti-jitter and anti-interference capabilities are crucial during the load-varying phase. This is because load fluctuations can cause system jitter and noise interference, and high jitter amplitudes can lead to unstable output. Therefore, during the load-varying phase, the observer needs to maintain a low jitter amplitude to ensure stable and accurate output.

[0123] In summary, the weighting is determined based on the importance of the observer's stability evaluation metrics during the load-variable phase. The higher weight for ACT ensures that the observer can quickly adapt to load changes and converge quickly. The remaining weights consider the observer's stability and anti-bounce capability under extreme conditions. This weighting helps evaluate and optimize the observer's performance during the load-variable phase.

[0124] Based on the above, the observer stability comprehensive evaluation matrix D = .

[0125] S03: Operate the fuel cell in different working modes, every Get the value of the accuracy index in the observer comprehensive evaluation index system, denoted as , u is taken as 1, 2, 3, 4; where: A1 is the root mean square error value, A2 is the mean absolute percentage error value, A3 is the mean absolute percentage error; A4 is the determination coefficient;

[0126] ;

[0127] ;

[0128] ;

[0129] ;

[0130] Where n represents the number of samples collected, m represents the position of each observation in the data set, Indicates the actual value, represents the observation value using the unscented Kalman filter observer, represents the predicted value of the observation using the unscented Kalman filter observer, Indicates the average value of the actual value. k indicates the number of times the data is collected; t k Indicates the time of the kth data collection; t k+1 Indicates the time of the data collected for the k+1th time;

[0131] Operate the fuel cell in different working modes, Get the value of the stability index in the observer comprehensive evaluation index system, denoted as , z is taken as 1, 2, 3, 4; where: X1 is the maximum convergence time; X2 is the average convergence time; X3 is the maximum jitter amplitude; X4 is the average jitter amplitude;

[0132] The numerical calculation formula of the stability index is as follows:

[0133] ;

[0134] ;

[0135] ;

[0136] ;

[0137] in, represents the time taken for data convergence in the dth data collection, and e represents the total number of data collections; is the maximum value of the observation value using the unscented Kalman filter observer, is the minimum value of the observation using the unscented Kalman filter observer.

[0138] The values ​​of sub-indicators in different working modes and X z Perform homogenization to obtain the accuracy normalized value Normalized value of the stability sub-index , specifically:

[0139] Normalized value of sub-indicators of accuracy ;

[0140] Normalized value of stability sub-index .

[0141] The accuracy sub-indicators under the comprehensive evaluation index system of the observer are obtained in three stages respectively. The superscripts 1, 2 and 3 represent the three stages, where 1 represents the startup stage, 2 represents the steady-state stage, and 3 represents the load-varying stage. It represents the value after the RMS value is normalized after using the observer in the startup phase; the details are as follows:

[0142] , , ;

[0143] The accuracy index normalization matrix is ​​constructed by normalizing the sub-indicators of accuracy : ,matrix It is a 3X4 matrix.

[0144] Similarly, the stability sub-indicators under the observer comprehensive evaluation index system are obtained in three stages, respectively:

[0145] , , ;

[0146] The stability index normalization matrix is ​​constructed by normalizing the sub-indices of stability : ,matrix It is a 3X4 matrix.

[0147] S04: The sub-indicator observer accuracy comprehensive evaluation matrix G, stability comprehensive evaluation matrix D and the obtained accuracy normalized matrix The matrix homogenized with the stability sub-index Conduct a comprehensive evaluation to determine the accuracy score of each sub-indicator under each working mode and stability score .Right now:

[0148] Determine the accuracy score of each sub-indicator under each working mode ;make , get a 3X4 matrix , where the matrix middle Denote it as the vth column of the matrix. When v=1, is the first column of the matrix; when v=2, is the second column of the matrix; when v=3, is the third column of the matrix; when v=4, is the fourth column of the matrix.

[0149] = = ;

[0150] Determine the stability score of each sub-indicator under each working mode .make , get a 3X4 matrix , where the matrix middle Denote it as the wth column of the matrix. When w=1, is the first column of the matrix; when w=2, is the second column of the matrix; when w=3, is the third column of the matrix; when w=4, is the fourth column of the matrix.

[0151] = = ;

[0152] S05: The accuracy score of each sub-index under each working mode obtained by the SMO observer is recorded as ; The accuracy index normalization matrix under each working mode is obtained by SMO observer and recorded as ; The stability score of each sub-index under each working mode obtained by the SMO observer is recorded as ; The stability index normalized matrix under each working mode is obtained by SMO observer and recorded as ;

[0153] The combined weighted-multiplication synthesis method is used to determine the accuracy sub-index score of the SMO observer in each working mode, that is, the accuracy score matrix of each sub-index under each working mode obtained by the SMO observer is and the observer accuracy normalized value matrix The combined weighted multiplication synthesis method is used for calculation, and the calculation results are shown in Table 1.

[0154] The combined weighted-multiplication synthesis method is used to determine the stability sub-index score of the SMO observer in each working mode, that is, the stability score matrix of each sub-index in each working mode obtained by the SMO observer is and the observer stability normalization matrix The calculation is performed using the combined weighting-multiplication synthesis method. The calculation results are shown in Table 2.

[0155] The accuracy score of each sub-indicator under each working mode obtained by the UKF observer is recorded as , the accuracy index normalized matrix under each working mode is obtained through the UKF observer and recorded as ; The stability index normalized matrix under each working mode is obtained by UKF observer and recorded as ;

[0156] The combined weighted-multiplication synthesis method is used to determine the accuracy sub-index score of the UKF observer in each working mode, that is, the accuracy score matrix of each sub-index under each working mode obtained by the UKF observer is and the observer accuracy normalization matrix The calculation is performed using the combined weighting-multiplication synthesis method. The calculation results are shown in Table 1.

[0157] The combined weighted-multiplication synthesis method is used to determine the stability sub-index score of the UKF observer in each working mode, that is, the stability score matrix of each sub-index in each working mode obtained by the UKF observer is and the observer stability normalization matrix The calculation is performed using the combined weighting-multiplication synthesis method. The calculation results are shown in Table 2.

[0158] Table 1: Accuracy sub-indicator scores

[0159]

[0160] Table , subscript 1 represents the root mean square error value of the accuracy sub-index; superscript 1 represents the startup phase, and SMO represents the accuracy score obtained using the SMO observer. , expressed as in the SMO observer The accuracy score of the u-th sub-indicator in the stage. , expressed as in the UKF observer The accuracy score of the u-th sub-indicator at stage.

[0161] Table 2: Stability sub-indicator scores

[0162]

[0163] , subscript 1 represents the maximum convergence time of the stability sub-index; superscript 1 represents the startup phase, and SMO represents the accuracy score obtained using the SMO observer. , expressed as in the SMO observer The stability score of the zth sub-indicator in the stage. , expressed as in the UKF observer The stability score of the zth sub-indicator in the stage.

[0164] Calculated Comprehensive scoring using SMO observer in the stage and comprehensive scoring using the UKF observer , the specific formula is as follows:

[0165] ;

[0166] ;

[0167] As shown in Table 3.

[0168] Table 3: Comprehensive score

[0169]

[0170] In the table, Represents the comprehensive score of the startup phase using the SMO observer.

[0171] S06: Establishing a fusion observer control strategy based on the comprehensive score F of each observer in each working mode.

[0172] exist In the stage, the comprehensive score of the state estimation using the SMO observer is compared Comprehensive score with state estimation using UKF observer ,like ,and , then in In the stage, the SMO observer is used as the observer for fuel cell state estimation to evaluate the performance; if ,and , then use the fusion observation strategy;

[0173] like ,and , then in In the stage, the UKF observer is used as the observer for fuel cell state estimation to evaluate the performance; if ,and , then the fusion observation strategy is used; a is the set value;

[0174] Use fusion observation strategies, such as Figure 2 As shown, the details are as follows:

[0175] The accuracy score and stability score using the SMO observer and UKF observer are calculated respectively, and the calculation formulas are as follows:

[0176] ;

[0177] ;

[0178] Where: For The accuracy score of the stage using the SMO observer; For The stability score of the stage using the SMO observer; For The accuracy score of the UKF observer was used in the stage; For The stability score of the UKF observer is used in the stage;

[0179] when > ,and > , then the SMO observer is used for performance evaluation;

[0180] when < ,and < , then the UKF observer is used for performance evaluation;

[0181] when > ,and < , the SMO observer is used for accuracy performance evaluation, and the UKF observer is used for stability performance evaluation;

[0182] when < ,and > , the UKF observer is used for accuracy performance evaluation, and the SMO observer is used for stability performance evaluation.

[0183] Example

[0184] S01: Working mode division: According to the change of working current, the working mode is divided into starting stage, steady state stage and variable load stage.

[0185] S02: Establish a comprehensive evaluation index system, including constructing the observer accuracy comprehensive evaluation matrix G and the observer stability comprehensive evaluation matrix D;

[0186] The comprehensive evaluation matrix G of the observer accuracy is:

[0187] G= ;

[0188] The comprehensive evaluation matrix D of observer stability is:

[0189] D= ;

[0190] S03: Use hydrogen supply methods under different working conditions to operate the fuel cell, every The values ​​of the accuracy index and the stability index in the observer comprehensive evaluation index system are obtained and normalized.

[0191] At a certain moment, the accuracy normalization matrix and stability normalization matrix of the SMO observer are obtained as follows:

[0192] ;

[0193] ;

[0194] At a certain moment, the accuracy normalization matrix and stability normalization matrix of the UKF observer are obtained as follows:

[0195] ;

[0196] ;

[0197] S04: Comprehensively evaluate the sub-indicator observer accuracy comprehensive evaluation matrix G, the stability comprehensive evaluation matrix D, the obtained accuracy normalized matrix J, and the stability sub-indicator normalized matrix H to determine the accuracy score of each sub-indicator under each working mode and stability score .

[0198] The accuracy score of each sub-indicator under each working mode obtained by the SMO observer and stability score They are:

[0199] = ;

[0200] = ;

[0201] Accuracy scores of each sub-indicator under each working mode obtained by UKF observer and stability score They are:

[0202] = ;

[0203] = ;

[0204] S05: Combined accuracy scores under each working mode , stability score The accuracy score, stability score and comprehensive score of the sub-indicators of each observer in each working mode are calculated by combining the weighted and multiplicative synthesis methods with the normalized values ​​of each observer.

[0205] The combined weighted-multiplication synthesis method is used to determine the accuracy sub-index score of the SMO observer in each working mode, that is, the accuracy score matrix of each sub-index under each working mode obtained by the SMO observer is and the observer accuracy normalized value matrix The calculation is performed using the combined weighting-multiplication synthesis method; the combined weighting-multiplication synthesis method is used to determine the accuracy sub-index score of the UKF observer under each working mode, that is, the accuracy score matrix of each sub-index under each working mode obtained by the UKF observer and the observer accuracy normalization matrix The calculation is performed using the combined weighting-multiplication synthesis method. The calculation results are shown in Table 4.

[0206] Table 4: Accuracy scores

[0207]

[0208] The combined weighted-multiplication synthesis method is used to determine the stability sub-index score of the SMO observer in each working mode, that is, the stability score matrix of each sub-index in each working mode obtained by the SMO observer is and the observer stability normalization matrix The combined weighted-multiplication synthesis method is used for calculation. The combined weighted-multiplication synthesis method is used to determine the stability sub-index score of the UKF observer under each working mode, that is, the stability score matrix of each sub-index under each working mode obtained by the UKF observer is and the observer stability normalization matrix The calculation is performed using the combined weighting-multiplication synthesis method. The calculation results are shown in Table 5.

[0209] Table 5: Stability scores

[0210]

[0211] Calculated Comprehensive scoring using SMO observer in the stage and comprehensive scoring using the UKF observer , as shown in Table 6.

[0212] Table 6: Comprehensive score

[0213]

[0214] S06: Establishing a fusion observer control strategy based on the comprehensive score F of each observer in each working mode.

[0215] During the startup phase, ,and , then the UKF observer is used to evaluate the performance during the startup phase;

[0216] In the steady-state phase, ,and , then use the fusion observation strategy:

[0217] The accuracy score and stability score using the SMO observer and UKF observer are calculated respectively, and the calculation formulas are as follows:

[0218] ;

[0219] ;

[0220] Right now > ,and < , the SMO observer is used for accuracy performance evaluation, and the UKF observer is used for stability performance evaluation;

[0221] During the load change phase, ,and , then use the fusion observation strategy:

[0222] ;

[0223] ;

[0224] Right now > ,and < , the SMO observer is used for accuracy performance evaluation, and the UKF observer is used for stability performance evaluation.

[0225] A fuel cell state evaluation device comprises a processor and a memory, wherein the memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the steps in the fuel cell state estimation method are executed.

[0226] A storage medium stores a computer program, which runs the steps of the fuel cell state estimation method when executed by a processor.

[0227] It should be understood that although this specification is described according to various embodiments, not every embodiment contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

[0228] The series of detailed descriptions listed above are only specific descriptions of feasible embodiments of the present invention. They are not intended to limit the scope of protection of the present invention. Any equivalent embodiments or changes that do not deviate from the technical spirit of the present invention should be included in the scope of protection of the present invention.

Claims

1. A fuel cell state estimation method, characterized in that: The steps include: Working mode division: According to the change of working current, the working mode is divided into starting stage, steady state stage and variable load stage; Establish a comprehensive evaluation index system, which includes an observer accuracy comprehensive evaluation matrix G and a stability comprehensive evaluation matrix D; the observer accuracy comprehensive evaluation matrix G includes the weights of accuracy indicators at different working stages, and the observer stability comprehensive evaluation matrix D includes the weights of stability indicators at different working stages; The fuel cell is operated at different stages to obtain and normalize the values ​​of the accuracy index and the stability index in the observer comprehensive evaluation index system; the values ​​of the accuracy index include the root mean square error value, the mean absolute percentage error value, the mean absolute percentage error, and the coefficient of determination; the values ​​of the stability index include the maximum convergence time, the average convergence time, the maximum jitter amplitude, and the average jitter amplitude; the normalized values ​​of the accuracy sub-indicators are used to form an accuracy index normalization matrix J, and the normalized values ​​of the stability sub-indicators are used to form a stability index normalization matrix H; The sub-indicator observer accuracy comprehensive evaluation matrix G, stability comprehensive evaluation matrix D and the obtained accuracy normalized matrix The matrix homogenized with the stability sub-index Conduct a comprehensive evaluation to determine the accuracy score of each sub-indicator under each working mode and stability score ; The accuracy sub-index score and stability sub-index score of the SMO observer under each working mode are determined by the combined weighting-multiplication synthesis method; the accuracy sub-index score and stability sub-index score of the UKF observer under each working mode are determined by the combined weighting-multiplication synthesis method; Determine the comprehensive scoring using SMO observer and the comprehensive scoring using UKF observer at different stages; The control strategy of the fusion observer is established based on the comprehensive score of each observer at different stages.

2. The fuel cell state estimation method according to claim 1, characterized in that: The comprehensive evaluation matrix G of observer accuracy is expressed as: , express The weight of the u-th accuracy indicator of the stage; , u=1 represents the root mean square error; u=2 represents the mean absolute percentage error; u=3 represents the maximum error; u=4 represents the coefficient of determination; , Indicates the startup phase, represents the steady-state phase, Indicates the load change stage; The comprehensive evaluation matrix D of observer stability is expressed as: , express The weight of the zth stability indicator of the stage; , z=1 represents the maximum convergence time; z=2 represents the average convergence time; z=3 represents the maximum jitter amplitude; z=4 represents the average jitter amplitude.

3. The fuel cell state estimation method according to claim 1, characterized in that: Run the fuel cell at different stages to obtain the accuracy index value in the observer comprehensive evaluation index system , , where: A1 is the root mean square error value, A2 is the mean absolute percentage error value, A3 is the mean absolute percentage error; A4 is the determination coefficient; ； ； ； ； Where n represents the number of samples collected, m represents the position of each observation in the data set, Indicates the actual value, represents the observation value using the unscented Kalman filter observer, represents the predicted value of the observation using the unscented Kalman filter observer, represents the average value of the actual value; k represents the number of times the data is collected; t k Indicates the time of the kth data collection; t k+1 Indicates the time of the data collected for the k+1th time; Run the fuel cell at different stages to obtain the value of the stability index in the observer comprehensive evaluation index system , Where: X1 is the maximum convergence time; X2 is the average convergence time; X3 is the maximum jitter amplitude; X4 is the average jitter amplitude; ； ； ； ； in, represents the time taken for data convergence in the dth data collection, and e represents the total number of data collections; is the maximum value of the observation value using the unscented Kalman filter observer, is the minimum value of the observation value using the unscented Kalman filter observer; The values ​​of sub-indicators at different stages and Perform homogenization to obtain the normalized values ​​of the accuracy sub-index to form the accuracy index normalized matrix J and the normalized values ​​of the stability sub-index to form the stability index normalized matrix ,in: ； ； Where: express The value of the u-th accuracy index after normalization using the observer in the stage; express The normalized value of the zth stability index after using the observer in the stage.

4. The fuel cell state estimation method according to claim 1, characterized in that: Determine the accuracy score of each sub-indicator under each working mode , ; Accuracy score of each sub-indicator in each working mode Expressed as , Represented as the vth column of the matrix; Determine the stability score of each sub-indicator under each working mode , ; Accuracy score of each sub-indicator in each working mode Expressed as , Represented as the w-th column of the matrix.

5. The fuel cell state estimation method according to claim 1, characterized in that: The accuracy sub-index score and stability sub-index score of the SMO observer under each working mode are determined by the combined weighting-multiplication synthesis method; the accuracy sub-index score and stability sub-index score of the UKF observer under each working mode are determined by the combined weighting-multiplication synthesis method. Specifically: The accuracy score of each sub-indicator under each working mode obtained by the SMO observer is recorded as ; The accuracy index normalization matrix under each working mode is obtained by SMO observer and recorded as ; The stability score of each sub-index under each working mode obtained by the SMO observer is recorded as ; The stability index normalized matrix under each working mode is obtained by SMO observer and recorded as ; The accuracy score matrix of each sub-indicator under each working mode obtained by the SMO observer and the observer accuracy normalized value matrix The combined weighted-multiplication synthesis method is used to calculate and obtain The accuracy score of the u-th sub-indicator in the stage is recorded as ; The stability score matrix of each sub-index under each working mode obtained by the SMO observer and the observer stability normalization matrix The combined weighted-multiplication synthesis method is used to calculate and obtain The stability score of the zth sub-index in the stage is recorded as ; The accuracy score of each sub-indicator under each working mode obtained by the UKF observer is recorded as , the accuracy index normalized matrix under each working mode is obtained through the UKF observer and recorded as ; The stability index normalized matrix under each working mode is obtained by UKF observer and recorded as ; The accuracy score matrix of each sub-indicator under each working mode obtained by the UKF observer and the observer accuracy normalization matrix The combined weighted-multiplication synthesis method is used to calculate and obtain the The accuracy score of the u-th sub-indicator in the stage is recorded as ; The stability score matrix of each sub-index under each working mode obtained by the UKF observer and the observer stability normalization matrix The combined weighted-multiplication synthesis method is used to calculate and obtain the The stability score of the zth sub-index in the stage is recorded as .

6. The fuel cell state estimation method according to claim 5, characterized in that: Determine the comprehensive scoring using the SMO observer and the comprehensive scoring using the UKF observer at different stages, specifically: According to the SMO observer The accuracy score of the u-th sub-indicator at stage and the stability score of the zth sub-index Sure SMO observer comprehensive score of the stage: ； According to the UKF observer The accuracy score of the u-th sub-indicator at stage and the stability score of the zth sub-index Sure UKF observer comprehensive score for the stage: ； Where: for The stage was scored comprehensively using the SMO observer; for The SMO observer was used to comprehensively score the stage.

7. The fuel cell state estimation method according to claim 1, characterized in that: The control strategy of the fusion observer is established based on the comprehensive score of each observer at different stages, specifically: exist In the stage, the comprehensive score of the state estimation using the SMO observer is compared Comprehensive score with state estimation using UKF observer , like ,and , then in In this stage, the SMO observer is used as the observer for fuel cell state estimation to perform performance evaluation; like ,and , then use the fusion observation strategy; like ,and , then in In this stage, the UKF observer is used as the observer for fuel cell state estimation to perform performance evaluation; like ,and , then the fusion observation strategy is used; a is the set value.

8. The fuel cell state estimation method according to claim 7, characterized in that: The fusion observation strategy is specifically as follows: The accuracy score and stability score using the SMO observer and UKF observer are calculated respectively, and the calculation formulas are as follows: ； ； Where: For The accuracy score of the stage was scored using the SMO observer; For The stability score of the stage using the SMO observer; For The accuracy score of the UKF observer was used in the stage; For The stability score of the UKF observer is used in the stage; when > ,and > , then the SMO observer is used for performance evaluation; when < ,and < , then the UKF observer is used for performance evaluation; when > ,and < , the SMO observer is used for accuracy performance evaluation, and the UKF observer is used for stability performance evaluation; when < ,and > , the UKF observer is used for accuracy performance evaluation, and the SMO observer is used for stability performance evaluation.

9. A state assessment device, characterized in that: The method comprises a processor and a memory, wherein the memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the steps of the fuel cell state estimation method according to any one of claims 1 to 8 are executed.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the fuel cell state estimation method according to any one of claims 1 to 8 are executed.

Images