Jul 2026· Journal of Marine Science and Technology· Vol 31, pp. 536 - 560· 0 citations· 37 references
TL;DR
A novel feature-engineering methodology is introduced that derives a physics-informed jitter variance from experiment–simulation discrepancies in time–frequency features, turning jittering from a heuristic into a principled step, and improves the ML usability of low-fidelity hydrodynamic time series.
Abstract
Machine learning (ML) and deep learning (DL) have significantly advanced maritime engineering and hydrodynamics by enabling data-driven modelling and prediction. In hydrodynamics, high-fidelity surrogates offer precision, but low-fidelity surrogates are often preferred for generating large training sets due to their low computational cost. Time series records are a central form of data across seakeeping, structural analysis, sea-state estimation, ship-response prediction, control, and manoeuvring; they encode both the system’s response to excitation and key vessel characteristics. To improve robustness, coverage, and perturbation resilience, augmentation techniques, such as jittering, scaling, warping, and permutation, are commonly applied, but typically tuned by trial-and-error and rarely grounded in physics. To address this, we introduce a novel feature-engineering methodology that derives a physics-informed jitter variance from experiment–simulation discrepancies in time–frequency features, turning jittering from a heuristic into a principled step. The method is evaluated on a spherical floating model tested in regular and irregular unidirectional waves and compared against numerical simulations. Beyond PDF/PSD-based representational fidelity comparison, the augmented datasets are further assessed through downstream simulation-to-experiment classification across benchmark ML/DL models. Compared with raw simulation and RMS-matched standard Gaussian jittering, the proposed physics-informed augmentation improved experimental test accuracy and macro-F1 across most model families. These results demonstrate that the proposed method is not merely a spectral noise adder, but a physics-guided augmentation strategy that improves the ML usability of low-fidelity hydrodynamic time series.
To address the critical scarcity of full life-cycle failure data for ball screws in industrial applications, this paper proposes the Hybrid Physics-Informed Data Augmentation (HPIDA) framework. This framework integrates an autoregressive (AR) system identification model with Archard’s wear law and Hertzian contact mechanics, constructing a comprehensive physical mapping chain from microscopic wear to macroscopic vibration signals. Furthermore, a domain randomization strategy is employed to synthesize high-fidelity, virtual full-life trajectories that encompass diverse degradation rates. Validation on the PredMAIN dataset demonstrates that the Wasserstein-1 distance for all generated signals strictly remains below 0.08, while both the power spectrum cosine similarity and the autocorrelation Pearson coefficient consistently exceed 0.9. The synthesized data exhibit a high degree of concordance with authentic signals across three dimensions: statistical distribution, frequency-domain structure, and temporal characteristics. Consequently, this approach provides robust, physically-grounded data to effectively support the training of deep learning-based Remaining Useful Life (RUL) prediction models under few-shot constraints.
The results demonstrate that PINN achieves more accurate and stable full-field vibration reconstructions than conventional PINNs, particularly under conditions involving high-frequency modes, and highlights the potential of hybrid data-physics neural frameworks as an efficient and reliable approach for solving complex PDE-governed dynamical systems.
Hai-Long Liu, S. Hedayatrasa, Yunpeng Zhu et al.· e-Journal of Nondestructive...· 0 citations
Industrial time-series signals, such as turbine temperature and rotational speed in aero-engines, are essential for monitoring the health and operational status of complex dynamical systems. However, collecting such data is often limited by harsh environments (e.g., high temperature and high pressure) and the high cost of experimental testing. To address this challenge, we introduce PhysDGM, a stepwise physics-embedded diffusion generative model for synthesizing time-series data that are consistent with the underlying physical laws of dynamical systems. PhysDGM embeds physical laws directly into each reverse diffusion step of the generative process, ensuring trajectory-level physical consistency, rather than enforcing constraints only at the final output. A large-scale AI-synthetic dataset (4.4 million samples, 20x scale-up) constructed by PhysDGM demonstrates strong fidelity across 34 datasets spanning turbofan engines, aero-engines, batteries, and chemical processes. After incorporating the synthetic data, the downstream task performance substantially surpassed that using real data alone by 48% for remaining useful life prediction, 15% for health indicator estimation, 22% for state-of-health assessment, and 20% for fault diagnosis. Moreover, it requires 10-20x less training data than existing approaches, substantially reducing the high cost of data collection in dynamical systems. We further demonstrate PhysDGM's potential in identifying early-stage faults in aero-engines by incorporating AI-synthesized data. In summary, PhysDGM provides a solid foundation for generating physically consistent industrial time-series, paving the way for expanding physics-guided AI into diverse data-scarce environments, including both industrial machinery and complex chemical reaction dynamics.
Haiteng Wang, Yunfei Zhu, Tao Wang et al.· 0 citations
Real-time simulation of tracked skid-steer vehicles is frequently bottlenecked by the computational cost of track– terrain interaction models. To address this, we replace the physics-based terramechanics function in a Simulink model with a neural network surrogate tailored for hardware-in-the-loop deployment. We train separate Multi-Layer Perceptrons (MLPs) for rigid (friction-limited) and soil (strength-limited) regimes to predict body-frame forces and yaw moment (Qx,Qy,Mz). Crucially, we deviate from standard offline training by enforcing input truthfulness: dataset generation relies strictly on logged block interfaces—including signal delays and solver artifacts—rather than idealized command profiles. To prevent temporal leakage, we utilize run-level data splitting and a custom weighted loss function that penalizes errors in the sensitive yaw moment channel. Finally, we demonstrate that a manual MEX deployment eliminates runtime overhead, reducing latency below the physics baseline to enable real-time execution.
Huseyin Sanli, M. Ö. Efe· International Conference on...· 0 citations
Deploying machine learning surrogates in scientific simulations faces multifaceted challenges, primary among which is the lack of Continual Learning (CL) capabilities—specifically, the inability to adapt to new physical regimes without significantly degrading performance on prior ones. This is particularly problematic for autoregressive surrogates of time-dependent Partial Differential Equations (PDEs), where small prediction errors can accumulate over long rollouts and new physical regimes overwrite previously learned dynamics. We formulate this adaptation as a CL problem, demonstrating that while standard Experience Replay (ER) is a robust baseline across Advection-Diffusion, Burgers’, and Navier-Stokes equations, storing full high-resolution rollouts can be memory-inefficient. To address this, we introduce Replay-TS, a temporal-slicing replay strategy that stores compact autoregressive windows sampled across past simulations. Through empirical analysis, we show that Replay-TS exploits the low-frequency spectral redundancy of physical systems to enable sparse supervision for rollout steps. By preserving the contiguous historical context and sparsely penalizing the autoregressive target steps, Replay-TS improves retention performance under a fixed memory budget by leveraging higher sample diversity. Replay-TS consistently outperforms standard ER methods across standard 1D and 2D streams, achieving over a 30% MSE reduction in a mixed-physics stream, while remaining architecture-agnostic.
Hamed Hemati, Binh Duong Nguyen, Stefan Sandfeld· Machine Learning for Computa...· 0 citations
This work proposes an efficient multilayer neural network framework to predict the pitching moment coefficient of canard-controlled missiles, significantly reducing the need for costly CFD data and providing a powerful surrogate tool for rapid design optimization based on trim angle of attack and geometric parameters.
M. Shojaeefard, Masoud Nobakhti· Proceedings of the Instituti...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
