Fatigue life prediction method and system for air compressor of fuel cell system

Abstract

The invention discloses a method and a system for predicting the fatigue life of an air compressor of a fuel cell system. The method comprises the following steps: determining a current damage value of the air compressor under a current time step length; determining a damage increment generated by the air compressor in the current time step according to the current damage value; accumulatively updating the current damage value according to the damage increment to obtain an updated damage value; and predicting the fatigue life of the air compressor according to the updated damage value. According to the method, high-precision and real-time dynamic prediction of the fatigue life of the air compressor of the fuel cell system under the coupling action of complex and changeable operation working conditions and a severe environment is realized, and the accuracy of life prediction is improved.

Description

Technical Field

[0001] This invention relates to the field of vehicle technology, and in particular to a method and system for predicting the fatigue life of an air compressor in a fuel cell system. Background Technology

[0002] Fuel cell stacks rely on high-pressure, clean air to maintain efficient electrochemical reactions. As a core component of the air supply system, the air compressor directly impacts system performance and reliability. Compared to passenger vehicles, the operating conditions of air compressors in fuel cell commercial vehicles are more demanding: on one hand, the high-power fuel cell stack causes the air compressor to operate at high speeds, inducing wide-band random vibrations; on the other hand, the low-frequency mechanical vibrations generated when the vehicle travels on complex road surfaces couple with the rotor's dynamic characteristics, exacerbating structural fatigue. Simultaneously, the alternating effects of low winter temperatures and high operating temperatures trigger thermomechanical fatigue, while the high humidity on the cathode side accelerates stress corrosion cracking of the aluminum alloy impeller. These multiple environmental loads significantly reduce the lifespan of critical components.

[0003] In related technologies, frequency-domain vibration fatigue analysis methods based on the Miner (linear cumulative damage criterion) are commonly used. However, this method assumes that fatigue damage is independent of the load sequence and that material properties are constant, making it difficult to accurately reflect the actual damage accumulation process of commercial vehicle air compressors under frequent changing operating conditions. In actual operation, the dynamic switching of working scenarios leads to strong time-varying characteristics of the impeller load spectrum, and damage exhibits path dependence. Furthermore, traditional stress amplitude-cycle count curves do not consider the degradation effect of material strength under humid and hot environments, resulting in significant deviations in life prediction results. Summary of the Invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art.

[0005] Therefore, one objective of this invention is to propose a method for predicting the fatigue life of air compressors in fuel cell systems. This method enables high-precision, real-time dynamic prediction of the fatigue life of air compressors in fuel cell systems under the coupled effects of complex and variable operating conditions and harsh environments, thereby improving the accuracy of life prediction.

[0006] Therefore, a second objective of this invention is to provide a system for predicting the fatigue life of an air compressor in a fuel cell system.

[0007] Therefore, a third objective of the present invention is to provide a vehicle.

[0008] To achieve the above objectives, a first aspect of the present invention discloses a method for predicting the fatigue life of an air compressor in a fuel cell system, comprising: determining the current damage value of the air compressor at a current time step; determining the damage increment generated by the air compressor within the current time step based on the current damage value; cumulatively updating the current damage value based on the damage increment to obtain an updated damage value; and predicting the fatigue life of the air compressor based on the updated damage value.

[0009] According to the fatigue life prediction method for air compressors in fuel cell systems of the present invention, after determining the current damage value of the air compressor at the current time step, the method calculates the damage increment within the current time step based on the damage value, and uses the damage increment to cumulatively update the current damage value to obtain the updated damage value. This process realizes the continuous transmission of damage state in time sequence, enabling the system to dynamically track the actual damage evolution path of the air compressor throughout its entire life cycle, ensuring the integrity and consistency of the fatigue accumulation process, and effectively avoiding evaluation deviations caused by damage state reset or discretization calculation. Subsequently, the fatigue life of the air compressor is predicted in real time based on the updated damage value, fully considering the cumulative effect of historical loads and the time-varying characteristics of actual operating conditions. This achieves high-precision, real-time dynamic prediction of the fatigue life of the air compressor in the fuel cell system under the coupled effects of complex and variable operating conditions and harsh environments, improving the accuracy of life prediction.

[0010] Furthermore, the fatigue life prediction method for air compressors in fuel cell systems according to the above embodiments of the present invention may also have the following additional technical features: In some embodiments, when accumulating and updating the current damage value based on the damage increment to obtain an updated damage value, the process includes: determining the current cycle number increment of the air compressor within the current time step; correcting the damage increment based on the current cycle number increment; and accumulating and updating the current damage value based on the corrected damage increment to obtain the updated damage value.

[0011] In some embodiments, the method for predicting the fatigue life of an air compressor in a fuel cell system further includes: adjusting the cycle number increment in the next time step based on the current damage value, the current cycle number increment, and a preset damage increment threshold, wherein the current cycle number increment is a pre-calibrated baseline increment in the initial calculation phase and an adaptive increment obtained from the adjustment of the previous time step in the non-initial calculation phase.

[0012] In some embodiments, predicting the fatigue life of the air compressor based on the updated damage value includes: when the updated damage value reaches a preset damage threshold, accumulating the incremental number of cycles corresponding to all executed time steps from the initial calculation stage to the target time step to obtain a cumulative number of cycles, wherein the target time step is the time step corresponding to the first time the updated damage value reaches the preset damage threshold; and determining the cumulative number of cycles as the fatigue life of the air compressor.

[0013] In some embodiments, determining the damage increment of the air compressor within the current time step based on the current damage value includes: acquiring the stress power spectral density, the reference resistance coefficient of the material used for the blades of the air compressor, and the current humidity and temperature of the material under the current operating environment; determining the humidity sensitivity coefficient, temperature sensitivity coefficient, and coupling coefficient for characterizing the effect of temperature and humidity interaction on material degradation using pre-calibrated simulation tests based on the preset target fatigue life, current temperature, and current humidity; determining an environmental correction value based on the humidity sensitivity coefficient, the temperature sensitivity coefficient, the coupling coefficient, the reference resistance coefficient, the current humidity, and the current temperature; and determining the damage increment based on the environmental correction value, the stress power spectral density, and the current damage value, combined with a pre-constructed nonlinear damage model.

[0014] In some embodiments, obtaining the benchmark resistance coefficient includes: determining the correlation data between the stress amplitude and the number of failure cycles corresponding to the material based on the material and a preset mapping relationship, wherein the mapping relationship includes the correspondence between multiple sets of materials, stress amplitudes and the number of failure cycles; based on the correlation data, using a preset fitting algorithm to fit the slope of the relationship curve formed by the stress amplitude and the number of failure cycles and the fatigue strength coefficient; and obtaining the benchmark resistance coefficient based on the slope and the fatigue strength coefficient.

[0015] In some embodiments, obtaining the stress power spectral density includes: obtaining stress time history data of the blade in a target region based on a first preset simulation model, and determining the power spectral density based on the stress time history data; obtaining the natural frequency and mode shape of the blade through a second preset simulation model, and determining the frequency response function based on the natural frequency and mode shape; inputting the frequency response function and the power spectral density into a preset algorithm model to obtain the stress power spectral density.

[0016] In some embodiments, determining the current damage value of the air compressor at the current time step includes: if the current time step is in the initial calculation stage of the air compressor, using a preset initial damage value as the current damage value; if the current time step is in the non-initial calculation stage of the air compressor, using the updated damage value obtained at the end of the previous calculation stage as the current damage value.

[0017] To achieve the above objectives, a second aspect of the present invention discloses a fatigue life prediction system for an air compressor in a fuel cell system, comprising: a first determining module for determining a current damage value of the air compressor at a current time step; a second determining module for determining a damage increment generated by the air compressor within the current time step based on the current damage value; an updating module for cumulatively updating the current damage value based on the damage increment to obtain an updated damage value; and a prediction module for predicting the fatigue life of the air compressor based on the updated damage value.

[0018] According to an embodiment of the present invention, a fatigue life prediction system for an air compressor in a fuel cell system comprises the following steps: First, a determining module determines the current damage value of the air compressor at the current time step. Second, a determining module calculates the damage increment within the current time step based on this damage value. An updating module then uses this damage increment to cumulatively update the current damage value, obtaining an updated damage value. This process achieves continuous temporal transmission of the damage state, enabling the system to dynamically track the actual damage evolution path of the air compressor throughout its entire lifespan. This ensures the integrity and consistency of the fatigue accumulation process and effectively avoids evaluation biases caused by damage state resets or discretized calculations. Subsequently, the prediction module performs real-time prediction of the air compressor's fatigue life based on the updated damage value, fully considering the cumulative effect of historical loads and the time-varying characteristics of actual operating conditions. This achieves high-precision, real-time dynamic prediction of the fatigue life of the air compressor in the fuel cell system under the coupled effects of complex and variable operating conditions and harsh environments, improving the accuracy of life prediction.

[0019] To achieve the above objectives, a third aspect of the present invention discloses a vehicle comprising: a fuel cell system; and a fatigue life prediction system for an air compressor of a fuel cell system as described in a second aspect of the present invention.

[0020] According to an embodiment of the present invention, after determining the current damage value of the air compressor at the current time step, the damage increment within the current time step is calculated based on the damage value, and the current damage value is cumulatively updated using the damage increment to obtain the updated damage value. This process realizes the continuous transmission of damage state in time sequence, enabling the system to dynamically track the actual damage evolution path of the air compressor throughout its entire life cycle, ensuring the integrity and consistency of the fatigue accumulation process, and effectively avoiding evaluation deviations caused by damage state reset or discretization calculation. Subsequently, the fatigue life of the air compressor is predicted in real time based on the updated damage value, fully considering the cumulative effect of historical loads and the time-varying characteristics of actual operating conditions. This achieves high-precision, real-time dynamic prediction of the fatigue life of the fuel cell system air compressor under the coupled effects of complex and variable operating conditions and harsh environments, improving the accuracy of life prediction.

[0021] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0022] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a flowchart of a method for predicting the fatigue life of an air compressor in a fuel cell system according to an embodiment of the present invention; Figure 2 This is a schematic diagram comparing experimental values ​​with fatigue life prediction results from a model according to an embodiment of the present invention; Figure 3 This is a curve showing the relationship between stress amplitude and failure cycle number according to an embodiment of the present invention; Figure 4 This is a structural block diagram of a fatigue life prediction system for an air compressor in a fuel cell system according to an embodiment of the present invention. Figure 5 This is a structural block diagram of a vehicle according to an embodiment of the present invention. Detailed Implementation

[0023] The embodiments of the present invention are described in detail below. The embodiments described with reference to the accompanying drawings are exemplary. The embodiments of the present invention are described in detail below.

[0024] The following is for reference. Figure 1 A method for predicting the fatigue life of an air compressor for a fuel cell system according to an embodiment of the present invention is described.

[0025] Figure 1This is a flowchart of a method for predicting the fatigue life of an air compressor in a fuel cell system according to an embodiment of the present invention. Figure 1 As shown, the method includes at least steps S1-S4.

[0026] Step S1: Determine the current damage value of the air compressor at the current time step.

[0027] In this embodiment, based on the historical damage status and current operating conditions of the air compressor, the current damage value of the air compressor at the current time step is obtained. The current damage value reflects the cumulative damage level of the air compressor from its initial use to the current time step and is a dynamic state variable of the damage evolution model. During algorithm execution, the current damage value is initialized or inherited depending on whether the system is running for the first time. For example, if the air compressor is being calculated for the first time, the current damage value at the current time step will be determined based on a preset initial damage value; if it is a subsequent run, the damage result at the end of the previous stage is inherited, thus ensuring the continuity of the damage accumulation process and consistency throughout the entire lifecycle, providing a baseline input for calculating the damage increment within the current time step.

[0028] Step S2: Determine the damage increment generated by the air compressor within the current time step based on the current damage value.

[0029] The damage increment is used to characterize the contribution of the current operating conditions to the fatigue degradation of the air compressor blades.

[0030] In this embodiment, determining the damage increment of the air compressor within the current time step based on the current damage value means that, based on the known cumulative damage state of the air compressor, and combined with the dynamic load conditions (such as stress power spectral density, temperature, humidity, and speed) within the current time step, the amount of fatigue damage added within the current time step is calculated using a preset model.

[0031] Among them, the current damage value serves as the initial reference state for the damage accumulation process, ensuring that the damage evolution has continuity and historical dependence; the calculation of damage increment can be based on frequency domain or time domain fatigue methods and consider the relationship between stress amplitude distribution and cycle number, or introduce environmental correction values ​​to reflect the degradation effect of actual operating conditions such as damp heat and corrosion on material properties.

[0032] Step S3: Update the current damage value cumulatively based on the damage increment to obtain the updated damage value.

[0033] In this embodiment, the damage increment calculated within the current time step is linearly added to the current damage value to obtain an updated value of the total cumulative fatigue damage at the end of the current time step, which is the updated damage value.

[0034] For example, let the current damage value be denoted as D. iThe damage increment is denoted as The updated damage value is denoted as D. 更新 Then we can get D. 更新 =D i + The updated damage values ​​can be used to characterize the latest damage status of critical components (such as blades) in the air compressor at the end of the current time step.

[0035] Furthermore, the updated damage value will be inherited as the current damage value for the next time step, realizing the continuous transfer of damage status between different time periods and ensuring the integrity and consistency of the fatigue evolution process throughout the entire life cycle. Through this cumulative update mechanism, the system can dynamically track the actual damage development path of the air compressor, avoiding evaluation deviations caused by state reset or discrete calculations, thereby improving the temporal continuity and engineering accuracy of fatigue life prediction.

[0036] Step S4: Predict the fatigue life of the air compressor based on the updated damage values.

[0037] In this embodiment, predicting the fatigue life of the air compressor based on the updated damage value refers to determining the damage evolution process based on the cumulative damage state obtained at the end of the current time step and calculating the remaining operating time or number of cycles the air compressor can withstand from the current moment to the failure moment, thereby achieving a dynamic assessment of its fatigue life. This process uses the updated damage value as a real-time characterization parameter of damage evolution. By analyzing the damage growth trend and the rate of change of the current damage increment, it estimates the number of subsequent time steps required to reach the failure threshold, and then accumulates the corresponding time or number of cycles to obtain the total life or remaining life. This prediction method fully considers the cumulative effect of historical loads and the time-varying characteristics of actual operating conditions, achieving high-precision, real-time dynamic prediction of the fatigue life of the air compressor in the fuel cell system under the coupled effects of complex and variable operating conditions and harsh environments, significantly improving the accuracy and engineering applicability of life assessment.

[0038] Therefore, in the embodiments of the present invention, after determining the current damage value of the air compressor at the current time step, the damage increment within the current time step is calculated based on the damage value, and the current damage value is cumulatively updated using the damage increment to obtain the updated damage value. This process realizes the continuous transmission of damage state in time sequence, enabling the system to dynamically track the actual damage evolution path of the air compressor throughout its entire life cycle, ensuring the integrity and consistency of the fatigue accumulation process, and effectively avoiding evaluation deviations caused by damage state reset or discretization calculation. Subsequently, the fatigue life of the air compressor is predicted in real time based on the updated damage value, fully considering the cumulative effect of historical loads and the time-varying characteristics of actual operating conditions. This achieves high-precision, real-time dynamic prediction of the fatigue life of the air compressor in the fuel cell system under the coupled effects of complex and variable operating conditions and harsh environments, improving the accuracy of life prediction.

[0039] In one embodiment of the present invention, when accumulating and updating the current damage value according to the damage increment to obtain the updated damage value, the method includes: determining the current cycle number increment of the air compressor within the current time step; correcting the damage increment according to the current cycle number increment; and accumulating and updating the current damage value based on the corrected damage increment to obtain the updated damage value.

[0040] In this embodiment, the current cycle increment of the air compressor within the current time step is first determined. This current cycle increment represents the number of effective stress cycles experienced under the current operating conditions (such as a specific speed and load spectrum). Subsequently, the initially calculated damage increment is corrected based on this current cycle increment to more accurately reflect the contribution of actual cyclic loads to fatigue damage. This correction process considers the nonlinear relationship between the cycle count and damage, avoiding errors caused by simple linear accumulation under variable loads or multi-axis vibrations. By introducing a correction mechanism for the cycle count increment, the damage increment depends not only on the stress level but also dynamically correlated with the actual number of cycles experienced, thereby improving the physical consistency and accuracy of damage calculation. Finally, the corrected damage increment is accumulated with the current damage value to complete the cumulative update of the damage state and obtain the updated damage value. This realistically simulates the fatigue evolution process of the air compressor under complex and variable operating conditions, thereby improving the accuracy of fatigue life prediction.

[0041] Specifically, the formula for calculating the damage increment in the non-resonant frequency range is: (1) (2) in, For the increase in damage, This is the corrected damage increment. P(S) represents the increment of the current cycle number, P(S) represents the stress power spectral density, M represents the environmental correction value, and D represents the environmental correction value. i is the current damage value, and S is the actual stress amplitude obtained in the experiment.

[0042] The formula for calculating the damage increment in the resonant frequency range is: (3) (4) Where E is the quality factor. Let be the nth natural frequency of the structure, that is, the frequency at which the system vibrates freely without external excitation; The half-power bandwidth, which can be calculated using the modal damping ratio, reflects the energy dissipation of the structure near the resonant frequency; that is, the larger the modal damping ratio, the larger the half-power bandwidth. A wider half-probability bandwidth indicates a faster energy decay in the resonance region, resulting in a lower quality factor E; conversely, a smaller half-probability bandwidth indicates a smaller modal damping ratio. The narrower the frequency selectivity, the greater the quality factor E, indicating that the system has better frequency selectivity and lower energy loss.

[0043] Therefore, in fatigue damage calculations, considering the phenomenon that the vibration response of a structure (such as the blades of an air compressor) is significantly amplified near the resonant frequency, it is necessary to differentiate the damage increment for different frequency bands. For non-resonant frequency bands far from the resonant region, the damage increment caused by stress power spectral density is calculated using the conventional model without additional correction; however, for frequency bands including the structure's nth natural frequency... Within the resonant frequency band, an amplification effect of the frequency response function needs to be introduced, based on the half-power bandwidth. (Determined by the modal damping ratio) The damage increment is amplified and corrected to accurately reflect the local stress concentration and fatigue acceleration caused by resonance. Since the load spectrum experienced by the air compressor in actual operation is mostly distributed in the non-resonance region, the damage contribution in most frequency bands is small and does not need to be amplified. The resonance correction mechanism is only activated when the excitation frequency is close to the natural frequency of key components such as blades, thus balancing computational efficiency and accuracy.

[0044] In the nonlinear damage model used in this application, such as the CDM (Continuum Damage Mechanics Model), damage evolution is described by a nonlinear differential equation, and its core mechanism is reflected in the denominator term (1 D i Above: As cumulative damage (i.e., current damage value D) increases... iWhen the denominator approaches 1 (i.e., the material is close to failure), the damage increment approaches zero, leading to a sharp increase in damage increment. This simulates the rapid propagation stage in the later stages of crack initiation, reflecting the nonlinear characteristics of damage acceleration. At the same time, the actual stress amplitude S directly reflects the influence of load intensity on the damage rate, while the environmental correction value M introduces the synergistic effect of the humid and hot environment on the degradation of blade material performance, reducing the material's load-bearing capacity and further accelerating damage accumulation. This enables the CDM model to more realistically simulate the entire fatigue failure process of an air compressor under the combined action of complex mechanical loads and harsh environments, thereby improving the physical consistency and engineering accuracy of life prediction.

[0045] In one embodiment of the present invention, the method for predicting the fatigue life of an air compressor in a fuel cell system further includes: adjusting the increment of the number of cycles in the next time step based on the current damage value, the current cycle increment, and a preset damage increment threshold, wherein the current cycle increment is a pre-calibrated baseline increment in the initial calculation stage and an adaptive increment obtained by adjustment from the previous time step in the non-initial calculation stage.

[0046] In the embodiment, during the initial calculation phase, the system has not yet established a damage evolution trend. Therefore, the baseline cycle number increment, which is pre-calibrated through simulation or experiment, is used as the initial input. For example, the pre-calibrated baseline increment is 100 to ensure stable startup of the calculation. After entering the non-initial calculation phase, the system intelligently adjusts the cycle number increment of the next time step based on the current damage value, the current cycle number increment, and the preset damage increment threshold.

[0047] Specifically, the increment of the current loop count The adjustment formula is: (5) in, This is the adjusted increment of the number of iterations. To preset the damage threshold, This is a smoothing factor.

[0048] Preset damage threshold For example, 10 -4 Preset damage threshold The smaller the range of values ​​for smoothing factor, the higher the calculation accuracy, but the more computational resources are consumed; The value range is, for example, (0.5, 1.0) to avoid abrupt changes in the time step.

[0049] In the specific calculation process, the criterion for judging high and low damage is based on the total cumulative damage. The calculation result is derived from the number of cycles under each operating condition. Corresponding failure cycle number The ratios are summed. For high damage cases, if the damage contribution at the current time step (i.e. (greater than the preset damage threshold) This indicates that the damage growth within the current time step is significant, requiring a more refined calculation step to capture the details of damage evolution; however, for low-damage cases, if the damage contribution of the current time step (i.e. The damage threshold is much smaller than the preset damage threshold. In such cases, the time step can be appropriately increased to improve computational efficiency. This adaptive adjustment mechanism allows for flexible control of the computational step size at different damage stages, ensuring precise tracking of damage development in high-risk areas and efficient computation in low-risk areas. This comprehensively improves the accuracy and real-time performance of fatigue life prediction for fuel cell system air compressors.

[0050] Specifically, the formula for calculating the total cumulative damage value is: (6) in, This represents the loop count for the current time step. This represents the number of failure loops at the current time step.

[0051] like > This indicates that the damage contribution is large at the current time step size, and the calculation needs to be refined to reduce the time step size.

[0052] like << This can increase the time step.

[0053] Therefore, the increment of the number of cycles gradually transforms from an initial fixed baseline value into an adaptive increment that dynamically adjusts as damage develops. This allows for efficient computation of the damage accumulation process in low-risk areas and precise tracking in high-risk areas, thereby optimizing the overall allocation of computational resources and improving the real-time performance and accuracy of fatigue life prediction.

[0054] In one embodiment of the present invention, predicting the fatigue life of an air compressor based on the updated damage value includes: when the updated damage value reaches a preset damage threshold, accumulating the incremental number of cycles corresponding to all executed time steps from the initial calculation stage to the target time step to obtain the cumulative number of cycles, wherein the target time step is the time step corresponding to the first time the updated damage value reaches the preset damage threshold; and determining the cumulative number of cycles as the fatigue life of the air compressor.

[0055] In an embodiment, when the updated damage value reaches the preset damage threshold for the first time, the incremental number of loops corresponding to all executed time steps from the initial calculation stage to the target time step is accumulated to obtain the cumulative number of loops, for example, denoted as N, where the target time step refers to the time step corresponding to the first time the updated damage value reaches the preset damage threshold.

[0056] Specifically, through the formula Calculate the cumulative number of iterations, where Let N represent the increment of the number of cycles within the i-th time step, and n be the total number of time steps from the initial calculation stage to the target time step. This process ensures that all cycles contributing to damage accumulation are included in the final fatigue life assessment. Finally, the calculated cumulative number of cycles N is determined as the fatigue life of the air compressor, thus achieving accurate life prediction based on the dynamic damage accumulation model.

[0057] In one embodiment of the present invention, determining the damage increment of the air compressor within the current time step based on the current damage value includes: obtaining the stress power spectral density, the reference resistance coefficient of the material used for the air compressor blades, and the current humidity and temperature of the material under the current operating environment; determining the humidity sensitivity coefficient, temperature sensitivity coefficient, and coupling coefficient used to characterize the effect of temperature and humidity interaction on material degradation based on the preset target fatigue life, current temperature, and current humidity using pre-calibrated simulation tests; determining an environmental correction value based on the humidity sensitivity coefficient, temperature sensitivity coefficient, coupling coefficient, reference resistance coefficient, current humidity, and current temperature; and determining the damage increment based on the environmental correction value, stress power spectral density, and current damage value, combined with a pre-constructed nonlinear damage model.

[0058] In this embodiment, the stress power spectral density P(S) borne by the air compressor blades, the baseline resistance coefficient of the blade material, and the current temperature and humidity parameters under the current operating environment are first obtained. Then, based on the preset target fatigue life and the current temperature and humidity, the material's response characteristics to environmental factors are determined through a pre-calibrated simulation test, such as a CFD (Computational Fluid Dynamics Simulation Model). These characteristics include the humidity sensitivity coefficient, the temperature sensitivity coefficient, and the coupling coefficient characterizing the interaction between temperature and humidity on material degradation. Then, the humidity sensitivity coefficient, the temperature sensitivity coefficient, and the coupling coefficient are combined with the baseline resistance coefficient, the current temperature, and the current humidity to calculate an environmental correction value. This environmental correction value is used to quantify the degradation effect of the humid and hot environment on the fatigue performance of the material. This environmental correction value is then introduced into a pre-constructed nonlinear damage model (such as CDM) and used together with the stress power spectral density and the current damage value as model inputs. Taking into account the load spectrum characteristics, the nonlinear damage evolution behavior of the material, and the environmental degradation effect, the damage increment within the current time step is finally calculated. This method overcomes the limitation of traditional fatigue models that ignore the time-varying influence of environmental factors, and realizes high-precision dynamic assessment of fatigue damage of air compressor blades under complex and variable operating conditions, thereby improving the accuracy of fatigue life prediction.

[0059] In the specific calculation process, it can be determined according to the following formula, namely: Let the above formula (2) be denoted as ,Right now: (7) Where m is a preset material damage index, which controls the damage accumulation rate. The larger m is, the more nonlinear the damage growth. Let M be the environmental correction value, and the formula for calculating the environmental correction value M is: (8) Where T is the current temperature, H is the current humidity, and K is the baseline resistance coefficient. Humidity sensitivity coefficient For temperature sensitivity coefficient, is the coupling coefficient.

[0060] Based on the preset target fatigue life d and formulas (7) and (8), the following relationship can be obtained: (9) Taking the logarithm of formula (9) yields the following relationship: (10) Wherein, U is a constant term related to stress conditions, number of cycles, and reference resistance coefficient K, the value of which is determined by the material fatigue characteristics and load state; V is an experimental error term, which can be obtained through... Figure 2 The comparison curves between the experimental values ​​and the fatigue life prediction results of the model are fitted to determine the deviation between the actual fatigue life and the theoretical calculation value.

[0061] For the identified air compressor turbine, to verify the accuracy of the life prediction method based on the CDM model proposed in this application, the traditional Miner linear cumulative damage theory was selected as a comparison benchmark. Multiphysics simulation was used to calculate the surface stress distribution of the turbine blades at different power levels, and fatigue life prediction was performed using both the Miner method and the CDM model proposed in this application. As shown in Table 1, the experimental calibration was conducted under five typical speed conditions (corresponding to five power levels), and environmental parameters such as temperature and humidity were collected under each condition and used as input conditions for the CDM model. Based on the constructed nonlinear damage evolution equation, combined with the stress power spectral density output by the Dirlik algorithm and the environmental correction term, the cumulative number of cycles required for the blades to reach the failure threshold at each power level was calculated. The results are shown in Table 1. Figure 2 As shown.

[0062] The predicted lifespans of the CDM model and the Miner method were compared with the fatigue lifespans measured in actual tests, and the relative errors were calculated. Error analysis shows that, across the full power range, the prediction results of the CDM model are closer to the experimental lifespans, and the relative error is significantly lower than that of the Miner method, especially under complex operating conditions such as variable load, high humidity, and high temperature. These results verify the effectiveness of the proposed method in considering environmental degradation, nonlinear damage accumulation, and history dependence, demonstrating that it has higher prediction accuracy and engineering applicability compared to traditional methods, and can more realistically reflect the fatigue evolution process of air compressor turbines in actual service environments.

[0063] Collected environmental data ( , Substituting into formula (10), we can obtain the following relationship: (11) in, The preset target fatigue life is the value corresponding to the real-time temperature and humidity under the current environment. This is the real-time temperature detected under the current environment. This represents the real-time humidity detected under the current environmental conditions. , It can be obtained by querying Table 1.

[0064] Rearranging equation (11) into matrix form, we get: (12) By solving the coefficient matrix of equation (12) using the least squares method, we can obtain: (13) Therefore, the humidity sensitivity coefficient can be calculated based on formula (13). Temperature sensitivity coefficient and coupling coefficient Then calculate the humidity sensitivity coefficient. Temperature sensitivity coefficient and coupling coefficient Substituting into equation (8) above, the environmental correction value M can be calculated.

[0065] Then, the environmental correction value M, stress power spectral density P(S), and current damage value are calculated. By combining a pre-built nonlinear damage model, such as CMD, and substituting it into the above formula (7), the damage increment can be calculated. .

[0066] Table 1 shows the operating conditions of the air compressor turbine blades.

[0067]

[0068] Table 1 In one embodiment of the present invention, obtaining the benchmark resistance coefficient includes: determining the correlation data between the stress amplitude and the number of failure cycles corresponding to the material based on the material and a preset mapping relationship, wherein the mapping relationship includes the correspondence between multiple sets of materials, stress amplitudes and the number of failure cycles; based on the correlation data, using a preset fitting algorithm to fit the slope of the relationship curve formed by the stress amplitude and the number of failure cycles and the fatigue strength coefficient; and obtaining the benchmark resistance coefficient based on the slope and the fatigue strength coefficient.

[0069] In this embodiment, firstly, based on the type of material used in the air compressor blades and in conjunction with a preset mapping relationship (i.e., Table 2), multiple sets of experimental or calibration data on the relationship between stress amplitude and failure cycle number for that material are determined. The mapping relationship stores fatigue life data for various materials at different stress levels, forming the correlation data of the relationship curve between stress amplitude and failure cycle number. Then, based on the obtained correlation data, a preset fitting algorithm (such as the least squares method) is used to perform curve fitting on the data to obtain the relationship curve between stress amplitude and failure cycle number describing the fatigue characteristics of the material. The slope of the curve (i.e., fatigue strength index) and the fatigue strength coefficient (i.e., the theoretical stress amplitude corresponding to one failure cycle, reflecting the fatigue resistance of the material) are extracted from the curve. Finally, the benchmark resistance coefficient used in the fatigue damage model is derived or calculated by combining the slope and the fatigue strength coefficient.

[0070] Table 2 shows the preset mapping relationships.

[0071]

[0072] Table 2 In a specific embodiment, the results can be obtained from Table 2. Figure 3 The curve showing the relationship between stress amplitude and failure cycle number is shown.

[0073] Use a pre-defined formula (e.g., the form of a Basquin equation), that is: (14) Taking the logarithm of equation (14) above, we get: (15) The slope k and fatigue strength coefficient C can be obtained by fitting using the least squares method, and then the reference resistance coefficient K≈ Based on empirical formulas Thus, the baseline resistance coefficient K is derived.

[0074] In one embodiment of the present invention, obtaining the stress power spectral density includes: obtaining stress time history data of the blade in the target area based on a first preset simulation model, and determining the power spectral density based on the stress time history data; obtaining the natural frequency and mode shape of the blade through a second preset simulation model, and determining the frequency response function based on the natural frequency and mode shape; inputting the frequency response function and the power spectral density into a preset algorithm model to obtain the stress power spectral density.

[0075] In this embodiment, firstly, based on a first preset simulation model, such as a multi-physics coupled transient dynamic finite element model, the dynamic response of the air compressor blade under typical operating conditions is simulated to obtain stress time history data of the blade in target areas (such as high-stress areas like the blade root and leading edge). Subsequently, frequency domain analysis is applied to this stress time history data, using fast Fourier transform combined with Welch power spectrum estimation technology to preliminarily calculate the power spectral density on the excitation side, thus characterizing the energy distribution characteristics of aerodynamic loads in the frequency domain. Simultaneously, using a second preset simulation model, such as a modal analysis model, the natural frequencies and corresponding mode shapes of the blade are extracted, and a frequency response function is constructed based on the modal superposition method to describe the dynamic response amplification characteristics of the structure under different frequency excitations, especially the response enhancement effect in the resonance region. Finally, the preliminarily obtained power spectral density and frequency response function are input into a preset algorithm model (such as the Dirlik algorithm), and the load-response transfer relationship is calculated based on the structural resonance characteristics and modal coupling, thereby obtaining the stress power spectral density reflecting the actual stress state of the blade. This method integrates transient simulation, modal analysis, and advanced spectral estimation techniques to achieve high-precision conversion from time-domain excitation to frequency-domain damage input, effectively improving the accuracy of load boundaries and engineering reliability in fatigue life prediction.

[0076] Specifically, the basic relationship of the frequency response function is: (16) in, Let n be the frequency response function for measuring the response at position j and applying excitation at position k; n is the modal order. and These are the mode shapes (modal vector components) of the r-th mode at points j and k, respectively. Let be the natural frequency of the r-th mode; Let be the damping loss factor of the r-th mode, where , Let be the damping ratio and be a known quantity.

[0077] The stress power spectral density P(S) is calculated using the Dirlik algorithm, which involves inputting the power spectral density and frequency response function into the Dirlik algorithm to calculate the stress power spectral density P(S).

[0078] Specifically, the power spectral density spectral moment is calculated, i.e.: (17) in, Let f be the power spectral density moment, f be the power spectral density, and G(f) be the frequency response function.

[0079] Calculate the Dirlik coefficient, i.e. , , Q, R, , .

[0080] in, The damage dominance factor for narrowband vibration is used to characterize the damage contribution weight of near-sinusoidal vibration. The larger the value, the more concentrated the power spectral density energy is at a few frequencies. It is a medium-amplitude cyclic correction factor; is the high-frequency small-cycle compensation factor; Q is the narrowband damage scale parameter, which determines the stress amplitude decay rate dominated by the narrowband component, and the smaller the value of Q, the higher the probability of large-amplitude cycles; R is the narrowband damage shape parameter, which is used to adjust the distribution width of medium-amplitude cycles and determines the flatness of the power spectral density. The first irregularity factor represents the randomness of the load, ranging from 0 to 1, where 1 indicates narrowband and 0 indicates wideband. The second irregularity factor reflects the concentration of power spectral density energy in the frequency domain; 1 indicates narrowband, and 0 indicates wideband. (And the first irregularity factor...) Together they determine the probability distribution shape of the stress amplitude.

[0081] in, , , Q, R, , The calculation formulas are as follows: (18) (19) (20) (twenty one) (twenty two) (twenty three) (twenty four) Finally, the stress power spectral density P(S) is obtained, that is: (25) In one embodiment of the present invention, determining the current damage value of the air compressor at the current time step includes: if the current time step is in the initial calculation stage of the air compressor, using the preset initial damage value as the current damage value; if the current time step is in the non-initial calculation stage of the air compressor, using the updated damage value obtained at the end of the previous calculation stage as the current damage value.

[0082] In this embodiment, if the current time step is in the initial calculation stage of air compressor fatigue life prediction, since there is no historical damage accumulation data, the system uses a preset initial damage value as the current damage value to initialize the damage calculation. The preset initial damage value is, for example, 0.02-0.05. When the current time step is in a non-initial calculation stage (i.e., the system has already performed at least one complete time step calculation), the system uses the updated damage value obtained at the end of the previous calculation stage as the current damage value for the current time step, thereby achieving continuous transmission and inheritance of the damage state between different time steps. This mechanism ensures the temporal continuity and historical dependence of the fatigue damage accumulation process, avoiding damage state reset due to equipment start-up / shutdown or calculation interruption, and enabling damage evolution to truly reflect the cumulative deterioration process of the air compressor throughout its service life. This design not only conforms to the fundamental physical law of irreversible fatigue damage but also provides an accurate and consistent state input basis for subsequent damage increment calculation, life prediction, and health status assessment based on the current damage value.

[0083] According to the fatigue life prediction method for air compressors in fuel cell systems of the present invention, after determining the current damage value of the air compressor at the current time step, the method calculates the damage increment within the current time step based on the damage value, and uses the damage increment to cumulatively update the current damage value to obtain the updated damage value. This process realizes the continuous transmission of damage state in time sequence, enabling the system to dynamically track the actual damage evolution path of the air compressor throughout its entire life cycle, ensuring the integrity and consistency of the fatigue accumulation process, and effectively avoiding evaluation deviations caused by damage state reset or discretization calculation. Subsequently, the fatigue life of the air compressor is predicted in real time based on the updated damage value, fully considering the cumulative effect of historical loads and the time-varying characteristics of actual operating conditions. This achieves high-precision, real-time dynamic prediction of the fatigue life of the air compressor in the fuel cell system under the coupled effects of complex and variable operating conditions and harsh environments, improving the accuracy of life prediction.

[0084] A further embodiment of the present invention discloses a fatigue life prediction system for an air compressor in a fuel cell system.

[0085] like Figure 4 As shown, the fatigue life prediction system 100 for the air compressor of a fuel cell system includes: a first determination module 110, a second determination module 120, an update module 130, and a prediction module 140.

[0086] The first determining module 110 is used to determine the current damage value of the air compressor at the current time step; the second determining module 120 is used to determine the damage increment generated by the air compressor within the current time step based on the current damage value; the updating module 130 is used to cumulatively update the current damage value based on the damage increment to obtain the updated damage value; and the prediction module 140 is used to predict the fatigue life of the air compressor based on the updated damage value.

[0087] In one embodiment of the present invention, when the update module 130 performs cumulative update on the current damage value based on the damage increment to obtain the updated damage value, it includes: determining the current cycle number increment of the air compressor within the current time step; correcting the damage increment based on the current cycle number increment; and cumulatively updating the current damage value based on the corrected damage increment to obtain the updated damage value.

[0088] In one embodiment of the present invention, the fatigue life prediction system 100 for the air compressor of a fuel cell system is further configured to: adjust the cycle number increment in the next time step according to the current damage value, the current cycle number increment and the preset damage increment threshold, wherein the current cycle number increment is a pre-calibrated baseline increment in the initial calculation stage and an adaptive increment obtained by the adjustment of the previous time step in the non-initial calculation stage.

[0089] In one embodiment of the present invention, when the prediction module 140 predicts the fatigue life of the air compressor based on the updated damage value, it includes: when the updated damage value reaches a preset damage threshold, accumulating the incremental number of loops corresponding to all executed time steps from the initial calculation stage to the target time step to obtain the cumulative number of loops, wherein the target time step is the time step corresponding to the first time the updated damage value reaches the preset damage threshold; and determining the cumulative number of loops as the fatigue life of the air compressor.

[0090] In one embodiment of the present invention, when the second determining module 120 determines the damage increment generated by the air compressor within the current time step based on the current damage value, it includes: acquiring the stress power spectral density, the reference resistance coefficient of the material used for the air compressor blades, and the current humidity and current temperature of the material under the current operating environment; determining the humidity sensitivity coefficient, temperature sensitivity coefficient, and coupling coefficient used to characterize the effect of temperature and humidity interaction on material degradation based on the preset target fatigue life, current temperature, and current humidity using a pre-calibrated simulation test; determining the environmental correction value based on the humidity sensitivity coefficient, temperature sensitivity coefficient, coupling coefficient, reference resistance coefficient, current humidity, and current temperature; and determining the damage increment based on the environmental correction value, stress power spectral density, and current damage value, combined with a pre-constructed nonlinear damage model.

[0091] In one embodiment of the present invention, when the second determining module 120 obtains the reference resistance coefficient, it includes: determining the correlation data between the stress amplitude and the number of failure cycles corresponding to the material according to the material and the preset mapping relationship, wherein the mapping relationship includes the correspondence between multiple sets of materials, stress amplitude and number of failure cycles; based on the correlation data, using a preset fitting algorithm to fit the slope of the relationship curve formed by the stress amplitude and the number of failure cycles and the fatigue strength coefficient; and obtaining the reference resistance coefficient according to the slope and the fatigue strength coefficient.

[0092] In one embodiment of the present invention, the second determining module 120, when acquiring the stress power spectral density, includes: acquiring stress time history data of the blade in the target area based on a first preset simulation model, and determining the power spectral density based on the stress time history data; acquiring the natural frequency and mode shape of the blade through the second preset simulation model, and determining the frequency response function based on the natural frequency and mode shape; inputting the frequency response function and the power spectral density into a preset algorithm model to obtain the stress power spectral density.

[0093] In one embodiment of the present invention, when the first determining module 110 determines the current damage value of the air compressor at the current time step, it includes: if the current time step is in the initial calculation stage of the air compressor, using the preset initial damage value as the current damage value; if the current time step is in the non-initial calculation stage of the air compressor, using the updated damage value obtained at the end of the previous calculation stage as the current damage value.

[0094] According to an embodiment of the present invention, a fatigue life prediction system 100 for an air compressor in a fuel cell system comprises a first determining module 110 determining the current damage value of the air compressor at the current time step, a second determining module 120 calculating the damage increment within the current time step based on the damage value, and an updating module 130 accumulating and updating the current damage value using the damage increment to obtain an updated damage value. This process realizes the continuous transmission of damage state in time sequence, enabling the system to dynamically track the actual damage evolution path of the air compressor throughout its entire life cycle, ensuring the integrity and consistency of the fatigue accumulation process, and effectively avoiding evaluation deviations caused by damage state reset or discretization calculation. Subsequently, a prediction module 140 predicts the fatigue life of the air compressor in real time based on the updated damage value, fully considering the cumulative effect of historical loads and the time-varying characteristics of actual operating conditions, thereby achieving high-precision, real-time dynamic prediction of the fatigue life of the air compressor in the fuel cell system under the coupled effects of complex and variable operating conditions and harsh environments, improving the accuracy of life prediction.

[0095] A further embodiment of the present invention discloses a vehicle.

[0096] In some embodiments, such as Figure 5 As shown, vehicle 200 includes fuel cell system 210 and fatigue life prediction system 100 for air compressor of fuel cell system as described in the above embodiments of the present invention.

[0097] According to the vehicle 200 of this embodiment, after determining the current damage value of the air compressor at the current time step, the damage increment within the current time step is calculated based on the damage value, and the current damage value is cumulatively updated using the damage increment to obtain the updated damage value. This process realizes the continuous transmission of damage state in time sequence, enabling the system to dynamically track the actual damage evolution path of the air compressor throughout its entire life cycle, ensuring the integrity and consistency of the fatigue accumulation process, and effectively avoiding evaluation deviations caused by damage state reset or discretization calculation. Subsequently, the fatigue life of the air compressor is predicted in real time based on the updated damage value, fully considering the cumulative effect of historical loads and the time-varying characteristics of actual operating conditions. This achieves high-precision, real-time dynamic prediction of the fatigue life of the fuel cell system air compressor under the coupled effects of complex and variable operating conditions and harsh environments, improving the accuracy of life prediction.

[0098] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example.

[0099] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for predicting the fatigue life of an air compressor in a fuel cell system, characterized in that, include: Determine the current damage value of the air compressor at the current time step; The damage increment generated by the air compressor within the current time step is determined based on the current damage value. The current damage value is cumulatively updated based on the damage increment to obtain the updated damage value; The fatigue life of the air compressor is predicted based on the updated damage values.

2. The method for predicting the fatigue life of an air compressor in a fuel cell system according to claim 1, characterized in that, When updating the current damage value based on the current damage increment to obtain the updated damage value, the process includes: Determine the increment of the current cycle number for the air compressor within the current time step; The damage increment is corrected based on the current cycle count increment, and the current damage value is cumulatively updated based on the corrected damage increment to obtain the updated damage value.

3. The method for predicting the fatigue life of an air compressor in a fuel cell system according to claim 2, characterized in that, Also includes: Based on the current damage value, the current cycle count increment, and the preset damage increment threshold, the cycle count increment in the next time step is adjusted. The current cycle count increment is a pre-calibrated baseline increment in the initial calculation phase and an adaptive increment obtained from the previous time step in the non-initial calculation phase.

4. The method for predicting the fatigue life of an air compressor in a fuel cell system according to claim 3, characterized in that, Predicting the fatigue life of the air compressor based on the updated damage values ​​includes: When the updated damage value reaches the preset damage threshold, the incremental number of loops corresponding to all executed time steps from the initial calculation stage to the target time step is accumulated to obtain the cumulative number of loops. The target time step is the time step corresponding to when the updated damage value first reaches the preset damage threshold. The cumulative number of cycles is determined as the fatigue life of the air compressor.

5. The method for predicting the fatigue life of an air compressor in a fuel cell system according to claim 1, characterized in that, Determining the damage increment of the air compressor within the current time step based on the current damage value includes: Obtain the stress power spectral density, the reference resistance coefficient of the material used for the blades of the air compressor, and the current humidity and temperature of the material under the current operating environment; Based on the preset target fatigue life, current temperature and current humidity, the humidity sensitivity coefficient, temperature sensitivity coefficient and coupling coefficient used to characterize the effect of temperature and humidity interaction on material degradation are determined by pre-calibrated simulation test. An environmental correction value is determined based on the humidity sensitivity coefficient, the temperature sensitivity coefficient, the coupling coefficient, the reference resistance coefficient, the current humidity, and the current temperature. The damage increment is determined based on the environmental correction value, the stress power spectral density, and the current damage value, combined with a pre-built nonlinear damage model.

6. The method for predicting the fatigue life of an air compressor in a fuel cell system according to claim 5, characterized in that, Obtaining the reference resistance coefficient includes: The correlation data of stress amplitude and failure cycle number corresponding to the material are determined based on the material and the preset mapping relationship, wherein the mapping relationship includes the correspondence between multiple sets of materials, stress amplitude and failure cycle number; Based on the aforementioned correlation data, a preset fitting algorithm is used to fit and obtain the slope of the relationship curve between the stress amplitude and the number of failure cycles, as well as the fatigue strength coefficient. The reference resistance coefficient is obtained based on the slope and the fatigue strength coefficient.

7. The method for predicting the fatigue life of an air compressor in a fuel cell system according to claim 5, characterized in that, Obtaining the stress power spectral density includes: The stress time history data of the blade in the target area is obtained based on the first preset simulation model, and the power spectral density is determined based on the stress time history data. The natural frequency and mode shape of the blade are obtained through a second preset simulation model, and the frequency response function is determined based on the natural frequency and mode shape. The frequency response function and the power spectral density are input into a preset algorithm model to obtain the stress power spectral density.

8. The method for predicting the fatigue life of an air compressor in a fuel cell system according to claim 1, characterized in that, Determining the current damage value of the air compressor at the current time step includes: If the current time step is in the initial calculation stage of the air compressor, the preset initial damage value will be used as the current damage value. If the current time step is in the non-initial calculation stage of the air compressor, the updated damage value obtained at the end of the previous calculation stage is used as the current damage value.

9. A fatigue life prediction system for an air compressor in a fuel cell system, characterized in that, include: The first determining module is used to determine the current damage value of the air compressor at the current time step; The second determining module is used to determine the damage increment generated by the air compressor within the current time step based on the current damage value. The update module is used to cumulatively update the current damage value based on the damage increment to obtain the updated damage value; The prediction module is used to predict the fatigue life of the air compressor based on the updated damage value.

10. A vehicle, characterized in that, include: Fuel cell system; as well as The fatigue life prediction system for air compressors in fuel cell systems as described in claim 9.

Images