Cost-Aware Multi-modal Multi-Fidelity Gaussian Process Fusion for Lithium-Ion Battery Pack Crash Damage Prediction

With the rapid development of new energy vehicles, fast and reliable prediction of power battery collision damage has become increasingly important. Traditional finite-element analysis is computationally expensive and difficult to deploy for rapid prediction under varying conditions. Although learning-based methods are faster, they usually rely on single-fidelity data: high-fidelity data is accurate but scarce and costly, while low-fidelity data is abundant but less reliable. Existing multi-fidelity methods alleviate this issue, yet often suffer from imbalanced sample allocation and weak cross-fidelity modeling. Moreover, current adaptive sampling strategies cannot dynamically determine the appropriate fidelity for different regions of the design space. To address these challenges, we propose HNGP-LCA, a multi-fidelity active learning framework for battery pack collision damage prediction. Our method consists of two components: (1) an Ensemble Nested Gaussian Process module that integrates single-layer and double-layer nested Gaussian process regression to better capture high–low fidelity correlations; and (2) a Location Information Cost-aware Active Learning strategy that leverages positional information to reconstruct expected improvement under different fidelities, enabling dynamic fidelity selection during sampling. Experiments on multiple synthetic benchmarks and a real battery pack engineering case demonstrate that HNGP-LCA achieves a better trade-off among accuracy, efficiency, and cost than strong baselines such as NARCO and MFBO. In the engineering case, it improves prediction accuracy by 0.6% over NARCO and 1.29% over MFBO, while reducing dependence on expensive high-fidelity data. These results show that HNGP-LCA provides an effective and practical solution for battery collision damage prediction.

In-vehicle 3D vision for perceiving dangerous driving behaviors

Modal metamodeling based uncertainty propagation for frequency response in non-proportional damping systems

Modal metamodeling based uncertainty propagation for frequency response in non-proportional damping systems

An Evidence Theory Based Decision Method for the Variable Sensitivity of Multi-Output Systems

A novel adaptive-weight ensemble surrogate model base on distance and mixture error

Surrogate models are commonly used as a substitute for the computation-intensive simulations in design optimization. However, building a high-accuracy surrogate model with limited samples remains a challenging task. In this paper, a novel adaptive-weight ensemble surrogate modeling method is proposed to address this challenge. Instead of using a single error metric, the proposed method takes into account the position of the prediction sample, the mixture error metric and the learning characteristics of the component surrogate models. The effectiveness of proposed ensemble models are tested on five highly nonlinear benchmark functions and a finite element model for the analysis of the frequency response of an automotive exhaust pipe. Comparative results demonstrate the effectiveness and promising potential of proposed method in achieving higher accuracy.

A mode tracking method in modal metamodeling for structures with clustered eigenvalues

Decentralized vibration control of smart constrained layer damping plate

The traditional centralized control strategy to vibration suppression of large-scale thin plate structures may increase the design difficulty of the controller. In this article, a decentralized vibration active control method is proposed to suppress the vibration of the thin plate structure with smart constrained layer damping treatment. First, the dynamics model of the smart constrained layer damping plate is established based on the finite element method, and the characteristics of viscoelastic materials with temperature and frequency are described by Golla-Hughes-McTavish damping model. Subsequently, a decentralized subsystem control model is obtained from the balanced model reduction method and complex mode truncation method. The modal test proves that the theoretical model is accurate. Then, the particular emphasis is placed on the stability and vibration attenuation of a decentralized system, which is composed of multiple subsystems. The local state feedback stabilization, using interaction of local state feedback and output feedback, is introduced to achieve system stability. To solve the practical problem of local state feedback, a decentralized controller with an observer is developed by adopting the pole placement method. Finally, the numerical simulation and hardware-in-the-loop experiment under different excitation are performed to investigate the effectiveness of decentralized vibration active control. The results demonstrate that the decentralized controller can effectively suppress the vibration, especially under mixed periodic signal and Gauss white noise signal.

Uncertainty propagation of frequency response functions using a multi-output Gaussian Process model

Numerical modeling and model updating for smart laminated structures with viscoelastic damping

Design Optimization of Vehicle Body NVH Performance Based on Dynamic Response Analysis

Active vibration control of thin-plate structures with partial SCLD treatment

A Stochastic Multivariate Validation Method for Dynamic Systems

As computer models become more powerful and popular, the complexity of input and output data raises new computational challenges. One of the key difficulties for model validation is to evaluate the quality of a computer model with multivariate, highly correlated and non-normal data, the direct application of traditional validation approaches does not appear to be suitable. This paper proposes a stochastic method to validate the dynamic systems. Firstly, a dimension reduction utilizing kernel principal component analysis (KPCA) is used to improve the computational efficiency. A probability model is then established by non-parametric kernel density estimation (KDE) method, and differences between the test data and simulation results are finally extracted to further comparative validation. This new approach resolves some critical drawbacks of the previous methods and improves the processing ability to nonlinear problem to validation the dynamic model. The proposed method and process are successfully illustrated through a real-world vehicle dynamic system example. The results demonstrate that the method of incorporate with KPCA and KDE is an effective approach to solve the dynamic model validation problem.