.png){kind=icon}
.png){kind=icon}
Return
Improved time-domain hybrid extrapolation method for vehicle durability load spectrum based on load component decomposition
.png){kind=icon}
PRE
AI
DOI：10.1016/j.measurement.2025.116660.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Time-domain Peak-over-Threshold (POT) extrapolation is a commonly used method for obtaining whole life- cycle load spectrum to assess the fatigue life of vehicle structures. However, it is insufficiently adaptive to random road loads with trend components, prone to extrapolation distortion. Additionally, the existing POT extrapolation method also suffers from unreasonable threshold selection, which may result in inadequate extremum fitting effects. To address these issues, based on the idea of 'decomposition-extrapolation-superposition', this paper proposes a time-domain hybrid extrapolation method for durability load spectra combined with load component decomposition. By analyzing the characteristics of road random load, it is separated into principal component and trend term loads using the Ensemble Empirical Mode Decomposition algorithm, so that the former is applicable for the POT extrapolation method. Additionally, a multi-criteria decision-making method based on multi-criteria weight analysis is proposed for the POT model threshold selection, enhancing the reasonableness of the threshold. Meanwhile, a time-domain reconstruction method based on Markov Chain Monte-Carlo simulation is proposed to reproduce the load variation intervals concerning the state transition characteristics of trend load. The complete extrapolated load spectrum is obtained by summing the extrapolated principal component and trend loads. Validation of the proposed method is conducted using load spectra. Results demonstrate that the proposed method effectively avoids distortion phenomenon caused by non-stationary load and reasonably extrapolates the load extremum. Moreover, compared to measured verification load spectrum, the parameter characteristics and fatigue damage of extrapolated load spectrum show good prediction performance.
Keywords:
Durability load spectrum
Time-domain hybrid extrapolation
Load component decomposition
POT model
Threshold selection
Multi-criteria decision-making
Markov Chain Monte-Carlo