Vehicle-mounted hydrogen fuel battery pack time-varying reliability analysis method and system
By analyzing data and identifying fault patterns of vehicle-mounted hydrogen fuel cell packs, a fault model is established, which solves the problem of insufficient fault identification and prediction in the prior art, and improves analysis accuracy and vehicle safety.
Patent Information
- Application Number
- CN202411005708.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-25
- Publication Date
- 2024-11-15
AI Technical Summary
The lack of identification and prediction of faults of vehicle-mounted hydrogen fuel cell packs in the prior art leads to low accuracy of reliability analysis results.
By collecting and cleaning the operating data of the vehicle-mounted hydrogen fuel cell pack, conducting statistical analysis and fault pattern identification, establishing a fault model to predict possible future failure conditions, and formulating a fault handling strategy based on the reliability evaluation results.
It realizes timely identification and prediction of faults of vehicle-mounted hydrogen fuel cell packs, improves the accuracy of reliability analysis, reduces maintenance costs and failure risks, and improves the safety and reliability of the vehicle.
Smart Images

Figure CN118967096A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data analysis, and in particular to a method and system for analyzing the time-varying reliability of a vehicle-mounted hydrogen fuel cell group. Background Art
[0002] The on-board hydrogen fuel cell group is a device that uses the redox reaction of hydrogen and oxygen to generate electricity. It consists of a hydrogen storage tank, an oxygen storage tank, a fuel cell stack, a battery control system, and a battery auxiliary system. The time-varying reliability analysis of the on-board hydrogen fuel cell group is to evaluate the reliability performance of the system in different time periods. Since the on-board hydrogen fuel cell group is a dynamically working device, its operating state will change over time. Therefore, time-varying reliability analysis can help us understand the reliability characteristics of the system in different time periods, including the changing trend of the failure rate, the probability distribution of failure occurrence, etc.
[0003] Prior art 1, application number: CN 202110329522.6 discloses a lithium-ion battery reliability assessment method based on a gray neural network model and a self-help method, and its prediction accuracy is significantly improved compared with the ordinary metabolic gray model and the BP neural network model. The combined model can make full use of the strong learning ability and good nonlinear mapping ability of the BP neural network model, as well as the characteristics of the simple operation of the gray model, and the characteristics of the self-help method with low sample quantity requirements, which can better fit the battery performance degradation curve and have a good extrapolation prediction effect. Although the pseudo-life of the battery when stored under different stresses is obtained, the reliability assessment is completed more efficiently and at a lower cost, and the final assessment result has strong objectivity and comprehensiveness. However, it can only evaluate the life, and cannot identify and predict faults, resulting in the inability to identify the reliability analysis results of the on-board hydrogen fuel cell.
[0004] Prior art 2, application number: CN 202211654042.8 discloses a reliability test method, device and computer-readable storage medium for lithium batteries. The reliability of lithium batteries is tested based on accelerated life tests, which can shorten the test time and test costs, and the reliability of lithium batteries is obtained based on the model between capacity retention rate and cycle number, which can obtain more accurate data, providing an effective basis for lithium battery quality assurance and replacement of spare parts. However, due to the different performance from lithium batteries, it is not suitable for vehicle-mounted hydrogen fuel cells, and accurate reliability analysis cannot be achieved.
[0005] Prior art three, application number: CN201310128348.4 discloses a dynamic reliability evaluation method for lithium-ion batteries based on polynomial fitting and life distribution, including obtaining the functional relationship between the operating condition conversion factor and the virtual failure threshold; obtaining the standard degradation trajectory of the product; obtaining the state offset of the product at the current moment; obtaining the state offset degree; and obtaining the dynamic real-time reliability of the product at each moment. Although it fully utilizes the information of similar products and the real-time monitoring information of the product to be evaluated, it realizes the dynamic reliability evaluation of lithium-ion batteries, and the evaluation effect is significant. However, the lack of identification of battery failures makes it impossible to accurately know the performance of the battery.
[0006] At present, the existing technologies 1, 2 and 3 lack the fault identification of hydrogen fuel cells and cannot predict future faults, resulting in low accuracy of reliability analysis results. Therefore, the present invention provides a method and system for time-varying reliability analysis of a vehicle-mounted hydrogen fuel cell group. By performing time-varying reliability analysis, the failure mode and reliability problems of the vehicle-mounted hydrogen fuel cell group can be discovered in time, providing a scientific basis for the normal operation and maintenance of the vehicle, improving the reliability and safety of the vehicle, and reducing maintenance costs and failure risks. Summary of the invention
[0007] The main purpose of the present invention is to provide a method and system for time-varying reliability analysis of a vehicle-mounted hydrogen fuel cell group, so as to solve the problem in the prior art of lack of fault identification of hydrogen fuel cells, inability to predict future failures, resulting in low accuracy of reliability analysis results.
[0008] To achieve the above objectives, in a first aspect, the present invention provides a method for analyzing the time-varying reliability of a vehicle-mounted hydrogen fuel cell group, the method comprising:
[0009] Collect the operating data of the on-board hydrogen fuel cell group, including working hours, working conditions, and fault record information; clean the collected data, including removing outliers and missing values; perform statistical analysis on the cleaned data, including calculating various indicators, fault frequency, fault duration, and fault mode, etc.;
[0010] By analyzing the fault data, different fault modes are identified, that is, the types of faults that occur in the on-board hydrogen fuel cell group under different working conditions; based on the existing fault data and models, fault prediction is performed to determine the possible fault conditions that may occur in the on-board hydrogen fuel cell group in future work;
[0011] Based on the fault data and prediction results, the reliability of the on-board hydrogen fuel cell group is evaluated, including the calculation of reliability indicators, mean time between failures and failure rate; based on the results of the reliability evaluation, corresponding fault handling strategies are formulated, including preventive maintenance and troubleshooting measures.
[0012] As a further improvement of the present invention, the process of statistically analyzing the cleaned data includes the following steps:
[0013] Perform descriptive statistical analysis on the cleaned data and calculate various indicators to describe the distribution and variation of the data, including mean, standard deviation, maximum, minimum and median;
[0014] Using the distribution and changes of various indicators, by setting a null hypothesis and an alternative hypothesis, we can determine whether the sample results are statistically significant at a certain confidence level. Hypothesis testing is used to determine whether an indicator is significantly different from an expected value.
[0015] Calculate confidence intervals, estimate the range of a certain indicator, calculate the confidence interval of the average value of the cleaned data, estimate the range of the overall average value, hypothesis test and confidence interval analysis, determine whether the cleaned data is statistically significant at a certain confidence level, and infer the characteristics of the population, and obtain the estimated value, confidence level and significance level information about the cleaned data population;
[0016] Among them, the failure frequency is calculated: according to the data after cleaning, the number of failures of the on-board hydrogen fuel cell group is counted, and it is divided by the total working time to obtain the failure frequency per unit time;
[0017] Calculate fault duration: Based on the cleaned data, record the start and end time of each fault, and calculate the time difference between the two time points to get the fault duration;
[0018] Analyze failure modes: Based on the cleaned data, identify different failure modes, that is, the types of failures that occur in the on-board hydrogen fuel cell group under different working conditions; by analyzing the failure records, find out the commonalities and laws of the failures to determine different failure modes;
[0019] Calculation of other reliability indicators: Based on the cleaned data, the mean time between failures and failure rate are also calculated.
[0020] As a further improvement of the present invention, the process of fault prediction includes the following steps:
[0021] Conduct statistical analysis on existing fault data, including fault frequency, duration, and fault mode; obtain the commonalities and rules of faults and identify different fault modes through analysis of fault data;
[0022] Based on the existing fault data, a fault model of the on-board hydrogen fuel cell group is established to describe the probability distribution of faults, the mechanism of fault occurrence, and the relationship between faults and working conditions;
[0023] The established fault model is used to predict possible faults that may occur in the future operation of the on-board hydrogen fuel cell group; based on the model and working conditions, the probability of fault occurrence or the expected number of faults is calculated.
[0024] As a further improvement of the present invention, the process of finding out different failure modes includes the following steps:
[0025] Conduct statistical analysis on the collected fault data, including calculating the frequency and duration of faults, obtaining fault patterns and trends, and finding out the commonalities and laws of faults;
[0026] Further analysis of the statistical analysis results was performed to identify different failure modes, which refer to the different types and causes of failures that occur in the on-board hydrogen fuel cell stack under different operating conditions;
[0027] According to the different fault modes identified, they are classified and described, including determining the fault type, fault cause and the mechanism of fault occurrence.
[0028] As a further improvement of the present invention, the process of establishing a fault model of a vehicle-mounted hydrogen fuel cell group includes the following steps:
[0029] Based on the existing fault data, feature engineering is performed and correlation analysis is used to extract fault-related features, including the probability distribution of faults, the mechanism of fault occurrence, and the relationship between faults and working conditions;
[0030] Select the gradient boosting ensemble model. According to the features and fault data, add weak learners iteratively. Each learner tries to correct the errors of the previous learner. During the training process, learn and adjust the parameters of the ensemble model according to the features and labels of the training set to obtain the fault model.
[0031] The accuracy, recall, and precision indicators are used to evaluate the performance of the trained fault model.
[0032] As a further improvement of the present invention, the process of extracting features related to faults using correlation analysis includes the following steps:
[0033] Collect fault data of the on-board hydrogen fuel cell group, including the time when the fault occurred and the type of fault;
[0034] For each fault sample, extract features related to the fault, including hydrogen flow, oxygen flow, temperature, pressure sensor data, and working conditions such as load and voltage;
[0035] The Pearson correlation coefficient of the correlation analysis method is used to calculate the correlation between each feature and the fault. Features with higher correlation are considered to be closely related to the fault. Based on the results of the correlation analysis, features that are strongly correlated with the fault are selected as input features in the training set.
[0036] As a further improvement of the present invention, the process of obtaining the fault model includes the following steps:
[0037] Initialize a weak learner of the decision tree, use the initialized weak learner to predict the training set, and calculate the difference between the actual value and the predicted value. The difference is called the residual.
[0038] Use gradient descent to update model parameters by calculating the negative gradient of the residual. Iteratively add more weak learners, each of which tries to correct the errors of the previous learner. Each iteration updates the model parameters.
[0039] According to the pre-set termination condition, the maximum number of iterations is reached, the iteration process is stopped, and the final integrated model is obtained.
[0040] As a further improvement of the present invention, the process of updating the model parameters using the gradient descent method comprises the following steps:
[0041] Calculate the predicted value of the current weak learner for the training set to obtain the output value of the current integrated model; calculate the difference between the actual value and the output value of the current integrated model, that is, the residual;
[0042] Calculate the negative gradient of the loss function based on the residual to update the model parameters;
[0043] Update the model parameters to reduce the loss function, and use gradient descent or other optimization algorithms to update the parameters; continue to iterate and add more weak learners, and each iteration attempts to correct the errors of the previous learners.
[0044] As a further improvement of the present invention, the process of evaluating the reliability of the vehicle-mounted hydrogen fuel cell group includes the following steps:
[0045] Based on the collected failure data, calculate the reliability index, mean time between failures and failure rate;
[0046] According to the fault prediction results and reliability indicators, the reliability level of the on-board hydrogen fuel cell group is evaluated; if the reliability indicator is low or there is a potential fault risk, the corresponding fault handling strategy is adopted;
[0047] Develop a fault handling strategy, including preventive maintenance and troubleshooting measures; preventive maintenance prevents faults from occurring through regular inspections and maintenance, and troubleshooting restores the normal operation of the equipment by repairing or replacing faulty parts.
[0048] In a second aspect, the present invention further provides a hydrogen-carrying fuel cell group time-varying reliability analysis system, which is used to implement the hydrogen-carrying fuel cell group time-varying reliability analysis method described in the first aspect, and the system comprises:
[0049] The data analysis module is responsible for collecting the operating data of the on-board hydrogen fuel cell group, including working hours, working conditions, and fault record information; cleaning the collected data, including removing outliers and missing values; and performing statistical analysis on the cleaned data, including calculating various indicators, fault frequency, fault duration, and fault mode;
[0050] The fault identification module is responsible for identifying different fault modes, that is, the types of faults that occur in the on-board hydrogen fuel cell group under different working conditions, by analyzing the fault data; and performing fault prediction based on the existing fault data and models to determine the possible faults that may occur in the on-board hydrogen fuel cell group in future work;
[0051] The result output module is responsible for evaluating the reliability of the on-board hydrogen fuel cell group based on the fault data and prediction results, including calculating the reliability index, mean time between failures and failure rate; and formulating corresponding fault handling strategies based on the results of the reliability evaluation, including preventive maintenance and troubleshooting measures.
[0052] The present invention collects and cleans the operating data of the on-board hydrogen fuel cell group, performs statistical analysis, obtains indicators such as the frequency, duration and mode of the fault, helps understand the fault situation of the on-board hydrogen fuel cell group, finds out the law and cause of the fault, and provides a data basis for subsequent fault prediction and reliability evaluation. By analyzing the fault data, identifying different fault modes, and performing fault prediction, it helps to predict possible faults in the future, and take measures in advance to prevent or deal with them; through the analysis and prediction of the fault mode, the design and operating condition control of the on-board hydrogen fuel cell group can be improved, and its reliability and stability can be improved. According to the fault data and the prediction results, the reliability of the on-board hydrogen fuel cell group is evaluated, and the corresponding fault handling strategy is formulated; through the reliability evaluation, the working state and performance of the on-board hydrogen fuel cell group can be evaluated, and its failure rate and mean time between failures within a certain period of time can be understood; according to the evaluation results, the corresponding preventive maintenance strategy can be formulated to improve the reliability and service life of the on-board hydrogen fuel cell group, and reduce the maintenance cost and the risk of failure. BRIEF DESCRIPTION OF THE DRAWINGS
[0053] Figure 1 A schematic diagram of the steps of an embodiment of a method for analyzing the time-varying reliability of a vehicle-mounted hydrogen fuel cell group according to the present invention;
[0054] Figure 2A schematic flow chart of the steps of performing statistical analysis on cleaned data in one embodiment of a method for analyzing the time-varying reliability of a vehicle-mounted hydrogen fuel cell group according to the present invention;
[0055] Figure 3 A schematic flow chart of the steps of fault prediction in one embodiment of the time-varying reliability analysis method for a vehicle-mounted hydrogen fuel cell group of the present invention;
[0056] Figure 4 A schematic flow chart of the steps of evaluating the reliability of a vehicle-mounted hydrogen fuel cell group according to an embodiment of a method for analyzing the time-varying reliability of a vehicle-mounted hydrogen fuel cell group of the present invention;
[0057] Figure 5 A schematic diagram of functional modules of an embodiment of a time-varying reliability analysis system for a vehicle-mounted hydrogen fuel cell group according to the present invention;
[0058] Figure 6 It is a structural schematic diagram of an embodiment of an electronic device of the present invention;
[0059] Figure 7 It is a schematic structural diagram of an embodiment of the storage medium of the present invention. DETAILED DESCRIPTION
[0060] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0061] The terms "first", "second" and "third" in the present invention are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined as "first", "second" and "third" can explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "multiple" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined. All directional indications (such as up, down, left, right, front, back...) in the embodiments of the present invention are only used to explain the relative position relationship, movement, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication also changes accordingly. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally also includes steps or units that are not listed, or optionally also includes other steps or units inherent to these processes, methods, products or devices.
[0062] Reference to an "embodiment" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiment may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
[0063] like Figure 1 As shown, this embodiment provides an embodiment of a time-varying reliability analysis method for a vehicle-mounted hydrogen fuel cell group. In this embodiment, the time-varying reliability analysis method for a vehicle-mounted hydrogen fuel cell group specifically includes the following steps:
[0064] Step S1: collecting the operating data of the on-board hydrogen fuel cell group, including working time, working conditions, fault records and other information; cleaning the collected data, including removing abnormal values and missing values; performing statistical analysis on the cleaned data, including calculating various indicators, such as fault frequency, fault duration and fault mode;
[0065] Step S2: by analyzing the fault data, identifying different fault modes, that is, the types of faults that occur in the on-board hydrogen fuel cell group under different working conditions; based on the existing fault data and models, performing fault prediction to determine the possible fault conditions that may occur in the on-board hydrogen fuel cell group in future work;
[0066] Step S3: Based on the fault data and prediction results, the reliability of the on-board hydrogen fuel cell group is evaluated, including calculating reliability indicators such as mean time between failures and failure rate; based on the results of the reliability evaluation, a corresponding fault handling strategy is formulated, including preventive maintenance and troubleshooting measures.
[0067] Preferably, step S1 of this embodiment collects and cleans the operating data of the on-board hydrogen fuel cell group, performs statistical analysis, obtains indicators such as the frequency, duration and mode of the fault, helps understand the fault situation of the on-board hydrogen fuel cell group, finds out the law and cause of the fault, and provides a data basis for subsequent fault prediction and reliability evaluation. Step S2 analyzes the fault data, identifies different fault modes, and performs fault prediction to help predict possible future faults and take measures to prevent or deal with them in advance; through the analysis and prediction of the fault mode, the design and working condition control of the on-board hydrogen fuel cell group can be improved to improve its reliability and stability. Step S3 evaluates the reliability of the on-board hydrogen fuel cell group based on the fault data and prediction results, and formulates corresponding fault handling strategies; through reliability evaluation, the working state and performance of the on-board hydrogen fuel cell group can be evaluated, and its failure rate and mean time between failures within a certain period of time can be understood; according to the evaluation results, corresponding preventive maintenance strategies can be formulated to improve the reliability and service life of the on-board hydrogen fuel cell group and reduce maintenance costs and failure risks.
[0068] In summary, the time-varying reliability analysis of the on-board hydrogen fuel cell group of this embodiment helps to understand and predict fault conditions, evaluate the reliability level, and formulate corresponding measures for prevention and treatment; improve its reliability and stability, thereby improving the safety, reliability and economy of the entire vehicle.
[0069] Furthermore, if Figure 2 As shown, the process of performing statistical analysis on the cleaned data in step S1 specifically includes the following steps:
[0070] Step S11: Perform descriptive statistical analysis on the cleaned data and calculate various indicators to describe the distribution and variation of the data, including mean, standard deviation, maximum, minimum and median;
[0071] Step S12: Using the distribution and changes of various indicators, by setting a null hypothesis and an alternative hypothesis, determine whether the sample results are statistically significant at a certain confidence level. Hypothesis testing is used to determine whether a certain indicator is significantly different from a certain expected value;
[0072] Step S13: Calculate the confidence interval, estimate the range of a certain indicator, calculate the confidence interval of the average value of the cleaned data, estimate the range of the overall average value, perform hypothesis testing and confidence interval analysis, determine whether the cleaned data is statistically significant at a certain confidence level, and infer the characteristics of the population, and obtain the estimated value, confidence level and significance level information of the cleaned data population;
[0073] Among them, the failure frequency is calculated: according to the data after cleaning, the number of failures of the on-board hydrogen fuel cell group is counted, and it is divided by the total working time to obtain the failure frequency per unit time;
[0074] Calculate fault duration: Based on the cleaned data, record the start and end time of each fault, and calculate the time difference between the two time points to get the fault duration;
[0075] Analyze failure modes: Based on the cleaned data, identify different failure modes, that is, the types of failures that occur in the on-board hydrogen fuel cell group under different working conditions; by analyzing the failure records, find out the commonalities and laws of the failures to determine different failure modes;
[0076] Calculation of other reliability indicators: Based on the cleaned data, other reliability indicators can also be calculated, such as mean time to failure (MTTF) and failure rate; MTTF represents the average time of trouble-free operation, which can be obtained by dividing the total working time by the number of failures; failure rate represents the probability of failure per unit time, which can be obtained by dividing the number of failures by the total working time.
[0077] Preferably, the descriptive statistical analysis of step S11 of this embodiment helps to understand the distribution and change of the cleaned data. By calculating indicators such as the mean, standard deviation, maximum value, minimum value and median, the central tendency, dispersion and distribution form of the data are obtained. The indicators can help understand the characteristics of the data and provide a basis for the subsequent inference of the overall characteristics. Step S12 uses the distribution and change of each indicator to determine whether the sample results are statistically significant at a certain confidence level through hypothesis testing, sets the null hypothesis and the alternative hypothesis, and compares the difference between the sample statistic and the expected value to determine whether the null hypothesis can be rejected, helping to determine whether a certain indicator is significantly different from the expected value, and inferring the overall characteristics. Step S13 calculates the confidence interval to estimate the range of a certain indicator. By calculating the confidence interval of the average value of the cleaned data, the range of the overall average value is estimated. Hypothesis testing and confidence interval analysis determine whether the cleaned data are statistically significant at a certain confidence level, infer the overall characteristics, and provide information about the estimated value, confidence level and significance level of the cleaned data population.
[0078] In summary, this embodiment identifies fault modes, evaluates fault frequency, duration and reliability indicators through the characteristics and overall features of the cleaned data, and provides a scientific basis for problem solving and decision-making; optimizes the operation and maintenance of the on-board hydrogen fuel cell group, improves its reliability and stability, and further enhances the performance and benefits of the entire system.
[0079] Furthermore, if Figure 3As shown, the fault prediction process in step S2 specifically includes the following steps:
[0080] Step S21: Statistically analyze the existing fault data, including the frequency, duration, and mode of the faults. By analyzing the fault data, obtain the commonalities and rules of the faults and find out different fault modes.
[0081] Step S22: Based on the existing fault data, a fault model of the on-board hydrogen fuel cell group is established to describe the probability distribution of the fault, the mechanism of the fault occurrence, and the relationship between the fault and the working conditions;
[0082] Step S23: using the established fault model, predicting possible faults of the on-board hydrogen fuel cell group in future operation; calculating the probability of fault occurrence or the expected number of faults based on the model and operating conditions.
[0083] Preferably, step S21 of this embodiment can obtain information such as the frequency, duration and failure mode of the fault by statistically analyzing the fault data, which is helpful to understand the fault conditions of the on-board hydrogen fuel cell group, find out the commonalities and laws of the faults, and provide a basis for establishing a fault model. Step S22 establishes a fault model of the on-board hydrogen fuel cell group to describe the probability distribution of the fault, the mechanism of the fault, and the relationship between the fault and the working conditions; by establishing a fault model, the cause and law of the fault can be better understood, providing a basis for fault prediction. Step S23 uses the established fault model to predict possible future faults; by calculating the probability or expected number of faults, possible faults can be predicted in advance, providing a basis for formulating preventive maintenance and fault handling strategies.
[0084] In summary, this embodiment improves the reliability and availability of the on-board hydrogen fuel cell group; by analyzing fault data and establishing a fault model, predicting possible fault conditions in the future, formulating corresponding preventive maintenance and fault handling strategies, the reliability of the on-board hydrogen fuel cell group is improved, the frequency and duration of faults are reduced, and the stability and safety of the system are improved.
[0085] Furthermore, the process of finding different fault modes in step S21 specifically includes the following steps:
[0086] Step S211: Statistically analyzing the collected fault data, including calculating the fault frequency, duration and other indicators, obtaining the fault occurrence pattern and trend, and finding out the commonalities and rules of the faults;
[0087] Step S212: further analyzing the results of the statistical analysis to identify different failure modes, where the failure modes refer to different failure types and causes of the on-board hydrogen fuel cell group under different working conditions;
[0088] Step S213: Classify and describe the different fault modes identified, including determining the fault type, fault cause, and fault occurrence mechanism.
[0089] Preferably, step S211 of this embodiment performs statistical analysis on the fault data, and can obtain indicators such as the frequency and duration of the fault, understand the mode and trend of the fault, help identify the commonality and laws of the fault, and provide a basis for subsequent fault prediction. Step S212 further analyzes the statistical results to identify different fault modes. Fault mode refers to the different types and causes of faults that occur in the on-board hydrogen fuel cell group under different working conditions; by identifying the fault mode, the characteristics and causes of the fault can be understood in more detail, providing guidance for subsequent fault prediction and fault handling. Step S213 classifies and describes the different fault modes identified, including determining the fault type, fault cause and mechanism of fault occurrence, etc., which can help establish a fault model, further describe the probability distribution of the fault, the mechanism of fault occurrence and the relationship between the fault and the working conditions; by establishing a fault model, future fault conditions can be predicted, so that corresponding measures can be taken in advance to reduce the risk of faults and ensure the normal operation of the on-board hydrogen fuel cell group.
[0090] In summary, this embodiment helps to understand and analyze fault data, identify fault modes, and establish fault models, thereby realizing the prediction of future fault conditions; predicting possible fault conditions in advance, taking corresponding preventive or maintenance measures, reducing the impact of faults on the operation of the on-board hydrogen fuel cell group, and improving safety and reliability. At the same time, fault prediction can also optimize maintenance plans and resource management, and improve maintenance efficiency and cost-effectiveness.
[0091] Furthermore, the process of establishing the fault model of the vehicle-mounted hydrogen fuel cell group in step S22 specifically includes the following steps:
[0092] Step S221: Perform feature engineering based on the existing fault data, and use correlation analysis to extract features related to the fault, including the probability distribution of the fault, the mechanism of the fault occurrence, and the relationship between the fault and the working condition;
[0093] Step S222: Select the gradient boosting integrated model, and iteratively add weak learners according to the features and fault data. Each learner attempts to correct the errors of the previous learner. During the training process, learn and adjust the parameters of the integrated model according to the features and labels of the training set to obtain the fault model.
[0094] Step S223: Use indicators such as accuracy, recall rate, and precision to evaluate the performance of the trained fault model.
[0095] Preferably, in step S221 of this embodiment, the feature engineering extracts features related to the fault through correlation analysis, including the probability distribution of the fault, the mechanism of the fault, and the relationship between the fault and the working conditions; a set of features that can describe and predict the fault is determined, which provides a basis for subsequent modeling. Step S222 selects a gradient boosting integrated model, and trains the fault model by iteratively adding weak learners. The gradient boosting algorithm is used to combine features and fault data to build a model that can accurately predict faults; the integrated model can integrate the prediction results of multiple weak learners to improve the accuracy and stability of the model. Step S223 evaluates the performance of the trained fault model through indicators such as accuracy, recall rate, and precision, objectively evaluates the fault model, and understands the prediction ability and reliability of the model. The evaluation results can guide further model optimization and improvement.
[0096] In summary, this embodiment can predict and diagnose the faults of the on-board hydrogen fuel cell group through characteristics and working conditions, which is of great significance for improving the safety and reliability of the vehicle and reducing the losses and maintenance costs caused by faults. In addition, the fault model can also provide a basis for formulating maintenance strategies and optimizing designs, thereby improving the overall performance of the on-board hydrogen fuel cell group.
[0097] Furthermore, the process of extracting features related to the fault using correlation analysis in step S221 specifically includes the following steps:
[0098] Step S2211: Collecting fault data of the vehicle-mounted hydrogen fuel cell group, including the time when the fault occurred, the type of fault, and other information;
[0099] Step S2212: for each fault sample, extract features related to the fault, including sensor data such as hydrogen flow, oxygen flow, temperature, pressure, and working conditions such as load, voltage, etc.;
[0100] Step S2213: Use the Pearson correlation coefficient of the correlation analysis method to calculate the correlation between each feature and the fault. Features with higher correlation are considered to be closely related to the fault. Based on the results of the correlation analysis, features that are strongly correlated with the fault are selected as input features in the training set.
[0101] Preferably, step S2211 of this embodiment obtains the fault data of the on-board hydrogen fuel cell group, providing a data basis for the subsequent establishment of a fault model. Step S2212 extracts features related to the fault, which can reflect the system conditions and states when the fault occurs, and provides input variables for the establishment of a fault model. Step S2213 determines the features closely related to the fault through a correlation analysis method, and selects the most representative and discriminative features as input variables, which can reduce the dimension of the features and improve the modeling effect and calculation efficiency of the fault model.
[0102] In summary, this embodiment can realize fault prediction and diagnosis of on-board hydrogen fuel cell groups by analyzing fault data and extracting relevant features to establish a fault model; it is of great significance to improve the reliability and safety of on-board hydrogen fuel cell groups, and can detect faults in time and take corresponding measures to improve the service life of the vehicle and user satisfaction.
[0103] Furthermore, the process of obtaining the fault model in step S222 specifically includes the following steps:
[0104] Step S2221: Initialize a weak learner of a decision tree, use the initialized weak learner to predict the training set, and calculate the difference between the actual value and the predicted value, the difference is called residual;
[0105] Step S2222: Use the gradient descent method to update the model parameters. The model parameters are updated by calculating the negative gradient of the residual. More weak learners are added iteratively. Each learner attempts to correct the errors of the previous learner. The model parameters are updated in each iteration.
[0106] Step S2223: According to the pre-set termination condition, the maximum number of iterations is reached, the iteration process is stopped, and the final integrated model is obtained.
[0107] Preferably, step S2221 of this embodiment obtains the initial prediction result and calculates the initial residual, which can provide a starting point for subsequent model training. Step S2222 continuously optimizes the model parameters through the gradient descent method so that the model can better fit the training data. Each iteration introduces a new weak learner and improves the performance of the overall model by correcting the errors of the previous learner. Step S2223 determines whether to continue iterating based on the pre-set termination conditions. If the maximum number of iterations or other set conditions are reached, the iteration is stopped to obtain the final integrated model; this integrated model is generated through multiple iterations and can be used for subsequent predictions and fault diagnosis.
[0108] In summary, this embodiment iteratively adds weak learners, continuously optimizes model parameters, and trains a powerful integrated model to identify and predict vehicle-mounted hydrogen fuel cell group faults, thereby improving the accuracy and efficiency of fault diagnosis. At the same time, this process can also help understand the mechanism of fault occurrence and the relationship between faults and working conditions, providing a reference for subsequent fault prevention and optimization.
[0109] Furthermore, the process of updating the model parameters using the gradient descent method in step S2222 specifically includes the following steps:
[0110] Step S22221: Calculate the predicted value of the current weak learner for the training set to obtain the output value of the current integrated model; calculate the difference between the actual value and the output value of the current integrated model, that is, the residual;
[0111] Step S22222: Calculate the negative gradient of the loss function based on the residual to update the model parameters;
[0112] Step S22223: Update the model parameters to reduce the loss function, and use the gradient descent method or other optimization algorithms to update the parameters; continue to iterate and add more weak learners, and each iteration attempts to correct the errors of the previous learner.
[0113] Preferably, step S22221 of this embodiment determines the prediction effect of the current model on the training set, and calculates the difference between the actual value and the predicted value, that is, the residual, which represents the error of the current model. Step S22222 calculates the negative gradient of the loss function based on the residual. The negative gradient represents the direction in which the loss function decreases the fastest. By using the negative gradient to update the model parameters, the model can move in the direction in which the loss function decreases. Step S22223 updates the model parameters using the gradient descent method or other optimization algorithms based on the calculated negative gradient, thereby reducing the loss function; by iteratively adding more weak learners and continuously updating the model parameters, the prediction ability of the model can be gradually improved and the errors of the previous learners can be corrected.
[0114] In summary, this embodiment continuously improves the prediction ability of the model through iteration; each learner attempts to correct the errors of the previous learner, and by updating the model parameters, the model can gradually fit the real rules in the training set, thereby obtaining more accurate prediction results; the final integrated model can be used to predict unknown data or solve practical problems.
[0115] Furthermore, if Figure 4 As shown, the process of evaluating the reliability of the vehicle-mounted hydrogen fuel cell group in step S3 specifically includes the following steps:
[0116] Step S31: Calculate reliability indicators such as mean time between failures and failure rate based on the collected failure data;
[0117] Step S32: According to the fault prediction result and in combination with the reliability index, the reliability level of the on-board hydrogen fuel cell group is evaluated; if the reliability index is low or there is a potential fault risk, a corresponding fault handling strategy is adopted;
[0118] Step S33: Formulate a fault handling strategy, including preventive maintenance and troubleshooting measures; preventive maintenance prevents the occurrence of faults through regular inspection and maintenance, and troubleshooting restores the normal operation of the equipment by repairing or replacing faulty parts.
[0119] Preferably, step S31 of this embodiment calculates reliability indicators, such as mean time between failures and failure rate, etc.; by statistically analyzing the fault data, the reliability indicators of the on-board hydrogen fuel cell group are obtained; the calculation of the mean time between failures can evaluate the stability and reliability of the equipment, and the calculation of the failure rate can quantify the probability of failure of the equipment in a unit time; it can provide important information about the performance and failure risk of the equipment. Step S32 evaluates the reliability level of the on-board hydrogen fuel cell group and combines the fault prediction results. By combining the fault prediction and reliability indicators, the reliability level of the on-board hydrogen fuel cell group is evaluated. If the reliability index is low or there is a potential risk of failure, the corresponding fault handling strategy can be adopted in advance to reduce the occurrence and impact of the failure. Evaluating the reliability level can help formulate reasonable maintenance and management measures to improve the reliability and service life of the equipment. Step S33 formulates a fault handling strategy, including preventive maintenance and troubleshooting measures; according to the evaluation results, formulate a corresponding fault handling strategy, preventive maintenance prevents the occurrence of faults through regular inspection and maintenance, can detect and repair potential problems in advance, thereby reducing the occurrence rate of faults, and troubleshooting restores the normal operation of the equipment by repairing or replacing faulty parts, which can quickly solve equipment failures and reduce downtime and production losses. Developing a fault handling strategy can improve equipment reliability and availability, and reduce maintenance costs and production risks.
[0120] In summary, this embodiment comprehensively evaluates the reliability level of the on-board hydrogen fuel cell group, predicts potential failure risks, and adopts corresponding failure handling strategies to ensure the normal operation of the equipment and extend its service life. Through effective reliability evaluation and failure handling measures, the reliability and stability of the equipment can be improved, the impact and losses caused by failures can be reduced, and the safety and reliability of the entire system can be improved.
[0121] like Figure 5 As shown, this embodiment also provides an embodiment of a time-varying reliability analysis system for a vehicle-mounted hydrogen fuel cell group. In this embodiment, the time-varying reliability analysis system for a vehicle-mounted hydrogen fuel cell group is used to implement the time-varying reliability analysis method for a vehicle-mounted hydrogen fuel cell group in the above embodiment. The system includes: a data analysis module 1, a fault identification module 2, and a result output module 3 that are electrically connected in sequence;
[0122] Among them, the data analysis module 1 is responsible for collecting the operating data of the on-board hydrogen fuel cell group, including working time, working conditions, fault records and other information; cleaning the collected data, including removing outliers and missing values; statistical analysis of the cleaned data, including calculating various indicators, such as fault frequency, fault duration and fault mode; the fault identification module 2 is responsible for identifying different fault modes, that is, the types of faults that occur in the on-board hydrogen fuel cell group under different working conditions by analyzing the fault data; fault prediction is performed based on the existing fault data and models to determine the possible fault conditions that may occur in the future work of the on-board hydrogen fuel cell group; the result output module 3 is responsible for evaluating the reliability of the on-board hydrogen fuel cell group based on the fault data and prediction results, including calculating reliability indicators, such as mean time between failures and failure rate; and formulating corresponding fault handling strategies based on the results of the reliability evaluation, including preventive maintenance and troubleshooting measures.
[0123] Preferably, the data analysis module 1 of this embodiment obtains indicators such as the frequency, duration and mode of fault occurrence by collecting, cleaning and statistically analyzing the operating data of the on-board hydrogen fuel cell group; understands the actual working conditions of the on-board hydrogen fuel cell group, and provides basic data for subsequent fault identification and reliability evaluation. The fault identification module 2 identifies different fault modes and performs fault prediction by analyzing the fault data; it can timely discover and predict possible faults of the on-board hydrogen fuel cell group, which helps to take maintenance measures in advance and avoid the impact of the fault on the normal operation of the vehicle. The result output module 3 evaluates the reliability of the on-board hydrogen fuel cell group based on the fault data and the prediction results, and calculates reliability indicators such as the mean time between failures and the failure rate; it can objectively evaluate the reliability level of the on-board hydrogen fuel cell group, and provide a basis for formulating reasonable maintenance strategies and improving the reliability of the vehicle.
[0124] Furthermore, the data analysis module 1 specifically includes:
[0125] The indicator calculation submodule is responsible for performing descriptive statistical analysis on the cleaned data and calculating various indicators to describe the distribution and changes of the data, including the mean, standard deviation, maximum, minimum and median values;
[0126] The hypothesis testing submodule is responsible for using the distribution and changes of various indicators to determine whether the sample results are statistically significant at a certain confidence level by setting a null hypothesis and an alternative hypothesis. The hypothesis test is used to determine whether a certain indicator is significantly different from a certain expected value.
[0127] The interval calculation submodule is responsible for calculating the confidence interval, estimating the range of a certain indicator, calculating the confidence interval of the average value of the cleaned data, estimating the range of the overall average value, hypothesis testing and confidence interval analysis, judging whether the cleaned data is statistically significant at a certain confidence level, and inferring the characteristics of the population, and obtaining the estimated value, confidence level and significance level information about the cleaned data population;
[0128] Among them, the failure frequency is calculated: according to the data after cleaning, the number of failures of the on-board hydrogen fuel cell group is counted, and it is divided by the total working time to obtain the failure frequency per unit time;
[0129] Calculate fault duration: Based on the cleaned data, record the start and end time of each fault, and calculate the time difference between the two time points to get the fault duration;
[0130] Analyze failure modes: Based on the cleaned data, identify different failure modes, that is, the types of failures that occur in the on-board hydrogen fuel cell group under different working conditions; by analyzing the failure records, find out the commonalities and laws of the failures to determine different failure modes;
[0131] Calculation of other reliability indicators: Based on the cleaned data, other reliability indicators can also be calculated, such as mean time to failure (MTTF) and failure rate; MTTF represents the average time of trouble-free operation, which can be obtained by dividing the total working time by the number of failures; failure rate represents the probability of failure per unit time, which can be obtained by dividing the number of failures by the total working time.
[0132] Preferably, the index calculation submodule of this embodiment can describe the distribution and change of data by calculating various indicators, provide a summary and description of the data, and help understand the characteristics and trends of the data. The hypothesis test submodule can determine whether the sample result is statistically significant, that is, whether there is a significant difference from the expected value, through hypothesis testing, to help determine whether the data has certain characteristics or changes. The interval calculation submodule can estimate the range of a certain indicator by calculating the confidence interval, and determine the significance of the data at a certain confidence level, provide estimated values and confidence level information about the data population, and help infer the characteristics of the population. Calculate the frequency of faults: By dividing the number of statistical faults by the total working time, the frequency of faults per unit time can be obtained, which helps to evaluate the fault situation and stability of the on-board hydrogen fuel cell group. Calculate the duration of the fault: By calculating the time difference between the start time and the end time of the fault, the duration of the fault can be obtained, which helps to understand the duration and impact range of the fault. Analyze the fault mode: By identifying different fault modes, the fault types that occur in the on-board hydrogen fuel cell group under different working conditions can be understood, which helps to determine the commonality and laws of the faults and provide a basis for fault handling and prevention. Calculation of other reliability indicators: By calculating indicators such as mean time between failures and failure rate, the reliability level of the on-board hydrogen fuel cell group can be evaluated to help determine the operating stability and failure risk of the equipment.
[0133] Furthermore, the fault identification module 2 specifically includes:
[0134] The pattern acquisition submodule is responsible for statistical analysis of existing fault data, including the frequency, duration, and mode of faults. By analyzing the fault data, the commonalities and rules of faults are obtained, and different fault modes are found.
[0135] The model building submodule is responsible for establishing the fault model of the on-board hydrogen fuel cell group based on the existing fault data, describing the probability distribution of the fault, the mechanism of the fault occurrence, and the relationship between the fault and the working conditions;
[0136] The model prediction submodule is responsible for using the established fault model to predict possible faults that may occur in the future operation of the on-board hydrogen fuel cell group; based on the model and working conditions, the probability of fault occurrence or the expected number of faults is calculated.
[0137] Preferably, the pattern acquisition submodule of this embodiment analyzes the existing fault data to obtain the commonalities and laws of the faults and find out different fault modes; it helps to understand the frequency, duration and characteristics of different fault modes, thereby providing a basis for subsequent fault prediction and fault management. The model construction submodule establishes a fault model based on the existing fault data to describe the probability distribution, occurrence mechanism and relationship between the fault and the working conditions; by establishing a fault model, the nature and laws of the fault can be deeply understood, and a scientific basis can be provided for fault prediction and fault management. The model prediction submodule uses the established fault model to predict possible fault conditions in the future; through the model and working conditions, the probability or expected number of faults can be calculated, which helps to discover potential fault risks in advance, take corresponding measures to prevent and deal with them, and ensure the reliability and safety of the on-board hydrogen fuel cell group.
[0138] In summary, this embodiment provides in-depth analysis and prediction capabilities for fault data, helping to better understand and manage fault problems of on-board hydrogen fuel cell groups. By identifying fault modes, establishing fault models, and performing fault prediction, the efficiency of fault diagnosis and repair can be improved, the impact of faults on vehicle operation and safety can be reduced, and the overall reliability and service life can be improved, while reducing maintenance costs and downtime.
[0139] Furthermore, the result output module 3 specifically includes:
[0140] The indicator calculation submodule is responsible for calculating reliability indicators such as mean time between failures and failure rate based on the collected fault data;
[0141] The level assessment submodule is responsible for assessing the reliability level of the on-board hydrogen fuel cell group based on the fault prediction results and reliability indicators; if the reliability indicator is low or there is a potential fault risk, the corresponding fault handling strategy is adopted;
[0142] The fault handling submodule is responsible for formulating fault handling strategies, including preventive maintenance and troubleshooting measures; preventive maintenance prevents the occurrence of faults through regular inspection and maintenance, and troubleshooting restores the normal operation of the equipment by repairing or replacing faulty parts.
[0143] Preferably, the index calculation submodule of this embodiment provides a quantitative assessment of the fault conditions of the on-board hydrogen fuel cell group by calculating the reliability index, which can accurately understand the frequency and failure rate of the fault, and provide a basis for subsequent fault handling. The technical effect of the level assessment submodule is to evaluate the reliability level of the on-board hydrogen fuel cell group by combining the fault prediction results and reliability indicators; it will help determine whether the on-board hydrogen fuel cell group meets the expected reliability requirements and promptly discover potential fault risks. The fault handling submodule formulates a fault handling strategy, including preventive maintenance and troubleshooting. Through regular inspection and maintenance, the occurrence of faults can be prevented and the reliability of the on-board hydrogen fuel cell group can be improved. Troubleshooting restores the normal operation of the equipment by repairing or replacing faulty parts, reducing the impact of faults on the performance and safety of the on-board hydrogen fuel cell group.
[0144] In summary, these submodules of this embodiment improve the reliability and safety of the on-board hydrogen fuel cell group. By analyzing and predicting fault data and formulating corresponding fault handling strategies, the downtime and maintenance costs caused by faults can be reduced, the availability and reliability of the on-board hydrogen fuel cell group can be improved, and its normal operation and work efficiency can be ensured.
[0145] like Figure 6 As shown, this embodiment provides an embodiment of an electronic device. In this embodiment, the electronic device 4 includes a processor 41 and a memory 42 coupled to the processor 41.
[0146] The memory 42 stores program instructions for implementing the time-varying reliability analysis method of the vehicle-mounted hydrogen fuel cell group according to any of the above-mentioned embodiments.
[0147] The processor 41 is used to execute program instructions stored in the memory 42 to perform time-varying reliability analysis of the on-vehicle hydrogen fuel cell group.
[0148] The processor 41 may also be referred to as a CPU (Central Processing Unit). The processor 41 may be an integrated circuit chip having signal processing capabilities. The processor 41 may also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.
[0149] Further, Figure 7The schematic diagram of the structure of the storage medium of an embodiment of the present application is that the storage medium 5 of the embodiment of the present application stores program instructions 51 that can implement all the above methods, wherein the program instructions 51 can be stored in the above storage medium in the form of a software product, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or terminal devices such as a computer, a server, a mobile phone, and a tablet.
[0150] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.
[0151] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit may be implemented in the form of hardware or in the form of software functional units. The above is only an implementation mode of the present invention, and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the present invention specification and drawings, or directly or indirectly used in other related technical fields, is also included in the patent protection scope of the present invention.
[0152] The specific implementation methods of the invention are described in detail above, but they are only examples, and the invention is not limited to the specific implementation methods described above. For those skilled in the art, any equivalent modification or substitution of the invention is also within the scope of the invention, therefore, the equalization, modification, improvement, etc. made without departing from the spirit and principle of the invention should be included in the scope of the invention.
Claims
1. A time-varying reliability analysis method for a vehicle-mounted hydrogen fuel cell group, characterized in that: The method comprises: Collect the operating data of the on-board hydrogen fuel cell group, including working hours, working conditions, and fault record information; clean the collected data, including removing outliers and missing values; perform statistical analysis on the cleaned data, including calculating various indicators, fault frequency, fault duration, and fault mode; By analyzing the fault data, different fault modes are identified, that is, the types of faults that occur in the on-board hydrogen fuel cell group under different working conditions; based on the existing fault data and models, fault prediction is performed to determine the possible fault conditions that may occur in the on-board hydrogen fuel cell group in future work; Based on the fault data and prediction results, the reliability of the on-board hydrogen fuel cell group is evaluated, including the calculation of reliability indicators, mean time between failures and failure rate; based on the results of the reliability evaluation, corresponding fault handling strategies are formulated, including preventive maintenance and troubleshooting measures.
2. The time-varying reliability analysis method of a vehicle-mounted hydrogen fuel cell group according to claim 1 is characterized in that: The process of statistical analysis of the cleaned data includes the following steps: Perform descriptive statistical analysis on the cleaned data and calculate various indicators to describe the distribution and variation of the data, including mean, standard deviation, maximum, minimum and median; Using the distribution and changes of various indicators, by setting a null hypothesis and an alternative hypothesis, we can determine whether the sample results are statistically significant at a certain confidence level. Hypothesis testing is used to determine whether an indicator is significantly different from an expected value. Calculate confidence intervals, estimate the range of a certain indicator, calculate the confidence interval of the average value of the cleaned data, estimate the range of the overall average value, hypothesis test and confidence interval analysis, determine whether the cleaned data is statistically significant at a certain confidence level, and infer the characteristics of the population, and obtain the estimated value, confidence level and significance level information about the cleaned data population; Among them, the failure frequency is calculated: according to the data after cleaning, the number of failures of the on-board hydrogen fuel cell group is counted, and it is divided by the total working time to obtain the failure frequency per unit time; Calculate fault duration: Based on the cleaned data, record the start and end time of each fault, and calculate the time difference between the two time points to get the fault duration; Analyze failure modes: Based on the cleaned data, identify different failure modes, that is, the types of failures that occur in the on-board hydrogen fuel cell group under different working conditions; by analyzing the failure records, find out the commonalities and laws of the failures to determine different failure modes; Calculation of other reliability indicators: Based on the cleaned data, the mean time between failures and failure rate are also calculated.
3. The time-varying reliability analysis method of a vehicle-mounted hydrogen fuel cell group according to claim 1 is characterized in that: The process of fault prediction includes the following steps: Conduct statistical analysis on existing fault data, including fault frequency, duration, and fault mode; obtain the commonalities and rules of faults and identify different fault modes through analysis of fault data; Based on the existing fault data, a fault model of the on-board hydrogen fuel cell group is established to describe the probability distribution of faults, the mechanism of fault occurrence, and the relationship between faults and working conditions; The established fault model is used to predict possible faults that may occur in the future operation of the on-board hydrogen fuel cell group; based on the model and working conditions, the probability of fault occurrence or the expected number of faults is calculated.
4. The time-varying reliability analysis method of a vehicle-mounted hydrogen fuel cell group according to claim 3 is characterized in that: The process of identifying different failure modes includes the following steps: Conduct statistical analysis on the collected fault data, including calculating the frequency and duration of faults, obtaining fault patterns and trends, and finding out the commonalities and laws of faults; Further analysis of the statistical analysis results was performed to identify different failure modes, which refer to the different types and causes of failures that occur in the on-board hydrogen fuel cell stack under different operating conditions; According to the different fault modes identified, they are classified and described, including determining the fault type, fault cause and the mechanism of fault occurrence.
5. The time-varying reliability analysis method of a vehicle-mounted hydrogen fuel cell group according to claim 3 is characterized in that: The process of establishing a fault model for a vehicle-mounted hydrogen fuel cell stack includes the following steps: Based on the existing fault data, feature engineering is performed and correlation analysis is used to extract fault-related features, including the probability distribution of faults, the mechanism of fault occurrence, and the relationship between faults and working conditions; Select the gradient boosting ensemble model. According to the features and fault data, add weak learners iteratively. Each learner tries to correct the errors of the previous learner. During the training process, learn and adjust the parameters of the ensemble model according to the features and labels of the training set to obtain the fault model. The accuracy, recall, and precision indicators are used to evaluate the performance of the trained fault model.
6. The time-varying reliability analysis method of a vehicle-mounted hydrogen fuel cell group according to claim 5 is characterized in that: The process of extracting features related to faults using correlation analysis includes the following steps: Collect fault data of the on-board hydrogen fuel cell group, including the time when the fault occurred and the type of fault; For each fault sample, extract features related to the fault, including hydrogen flow, oxygen flow, temperature, pressure sensor data, and working conditions such as load and voltage; The Pearson correlation coefficient of the correlation analysis method is used to calculate the correlation between each feature and the fault. Features with higher correlation are considered to be closely related to the fault. Based on the results of the correlation analysis, features that are strongly correlated with the fault are selected as input features in the training set.
7. The time-varying reliability analysis method of a vehicle-mounted hydrogen fuel cell group according to claim 5, characterized in that: The process of obtaining the fault model includes the following steps: Initialize a weak learner of the decision tree, use the initialized weak learner to predict the training set, and calculate the difference between the actual value and the predicted value. The difference is called the residual. Use gradient descent to update model parameters by calculating the negative gradient of the residual. Iteratively add more weak learners, each of which tries to correct the errors of the previous learner. Each iteration updates the model parameters. According to the pre-set termination condition, the maximum number of iterations is reached, the iteration process is stopped, and the final integrated model is obtained.
8. The time-varying reliability analysis method of a vehicle-mounted hydrogen fuel cell group according to claim 7 is characterized in that: The process of updating model parameters using gradient descent includes the following steps: Calculate the predicted value of the current weak learner for the training set to obtain the output value of the current integrated model; calculate the difference between the actual value and the output value of the current integrated model, that is, the residual; Calculate the negative gradient of the loss function based on the residual to update the model parameters; Update the model parameters to reduce the loss function, and use gradient descent or other optimization algorithms to update the parameters; continue to iterate and add more weak learners, and each iteration attempts to correct the errors of the previous learners.
9. The time-varying reliability analysis method of a vehicle-mounted hydrogen fuel cell group according to claim 1, characterized in that: The process of evaluating the reliability of a vehicle-mounted hydrogen fuel cell stack includes the following steps: Based on the collected failure data, calculate the reliability index, mean time between failures and failure rate; According to the fault prediction results and reliability indicators, the reliability level of the on-board hydrogen fuel cell group is evaluated; if the reliability indicator is low or there is a potential fault risk, the corresponding fault handling strategy is adopted; Develop a fault handling strategy, including preventive maintenance and troubleshooting measures; preventive maintenance prevents faults from occurring through regular inspections and maintenance, and troubleshooting restores the normal operation of the equipment by repairing or replacing faulty parts.
10. A hydrogen-carrying fuel cell group time-varying reliability analysis system, which is used to implement the hydrogen-carrying fuel cell group time-varying reliability analysis method according to any one of claims 1 to 9, characterized in that: The system comprises: The data analysis module is responsible for collecting the operating data of the on-board hydrogen fuel cell group, including working hours, working conditions, and fault record information; cleaning the collected data, including removing outliers and missing values; and performing statistical analysis on the cleaned data, including calculating various indicators, fault frequency, fault duration, and fault mode; The fault identification module is responsible for identifying different fault modes, that is, the types of faults that occur in the on-board hydrogen fuel cell group under different working conditions, by analyzing the fault data; and performing fault prediction based on the existing fault data and models to determine the possible faults that may occur in the on-board hydrogen fuel cell group in future work; The result output module is responsible for evaluating the reliability of the on-board hydrogen fuel cell group based on the fault data and prediction results, including calculating the reliability index, mean time between failures and failure rate; and formulating corresponding fault handling strategies based on the results of the reliability evaluation, including preventive maintenance and troubleshooting measures.
Citation Information
Cited By
Method and device for determining a time and a type of a maintenance measure of an energy converter system or energy storage system using semantic modelling
WO2026131522A1
