A Multi-Model Non-Intrusive Reduced-Order Framework for Parametric Erosion Prediction via Kinematic Cross-Moment Compression
Abstract
High-fidelity Eulerian--Lagrangian simulations of solid particle erosion in curved pipes require hours of compute per operating point, preventing rapid parameter sweeps and real-time wear assessment. Existing reduced-order models (ROMs) speed up these evaluations, yet they are typically trained on a single, fixed empirical erosion formula (e.g., Oka or Finnie). Changing the material law or target hardness then requires a complete retrain of the surrogate. Here, we present a non-intrusive reduced-order framework that avoids this model-locking by approximating the underlying particle collision kinematics instead of scalar wear rates. Specifically, we project and compress 23 Eulerian boundary cross-moments (𝔼[Vpu sinvαp coswαp]) across the pipe surface. Using 372 high-fidelity CFD-DPM cases of 90∘ elbows over three bend ratios (R/D ∈ {1.5, 2.0, 5.0}), five Reynolds numbers, five density ratios, and six particle diameters in the inertial regime (St > 1), we evaluate a hybrid compression scheme. Linear Proper Orthogonal Decomposition (POD) and Mode-1 tensor unfolding SVD are combined with block-wise Convolutional Autoencoders (CNN-AE) to handle both broad convective transport and localized impact craters. An anisotropic Gaussian Process Regression (GPR) surrogate maps four dimensionless Π-groups to the compressed latent space, evaluating full 2D wear topographies in roughly 2 ms (R2 > 0.99 on primary kinematic fields). Because kinematics are decoupled from material damage laws, the resulting surrogate evaluates multiple empirical models post-hoc exactly matching Finnie and closely approximating Oka, McLaury, and Arabnejad without retraining.
Related
- Non intrusive parametric Reduced Order Models based on physics for incompressible flow
- Efficient SPAD reconstruction framework with time-frequency domain compression and lightweight joint depth-intensity optimization
- Fisher Information Dynamics: A Kinematic Framework for Phase Space Ordering with Applications to Shock Layers and Turbulence
- A robust framework for the comprehensive characterisation of parametric down-conversion sources
- A Physics-Driven Framework for Parametric Periodic-Flow Modeling and Finite-Amplitude Aeroelastic Response Analysis
BibTeX
@misc{yadav2026multimodel,
title = {{A Multi-Model Non-Intrusive Reduced-Order Framework for Parametric Erosion Prediction via Kinematic Cross-Moment Compression}},
author = {Animesh Yadav and Rajesh Kumar Shukla and Ravinder Kumar Duvedi},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.09997},
eprint = {2609.09997},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.09997},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - A Multi-Model Non-Intrusive Reduced-Order Framework for Parametric Erosion Prediction via Kinematic Cross-Moment Compression
AU - Animesh Yadav
AU - Rajesh Kumar Shukla
AU - Ravinder Kumar Duvedi
PY - 2026
DA - 2026/09/09
PB - arXiv
DO - 10.48550/arxiv.2609.09997
UR - https://arxiv.org/abs/2609.09997
AB - High-fidelity Eulerian--Lagrangian simulations of solid particle erosion in curved pipes require hours of compute per operating point, preventing rapid parameter sweeps and real-time wear assessment. Existing reduced-order models (ROMs) speed up these evaluations, yet they are typically trained on a single, fixed empirical erosion formula (e.g., Oka or Finnie). Changing the material law or target hardness then requires a complete retrain of the surrogate. Here, we present a non-intrusive reduced-order framework that avoids this model-locking by approximating the underlying particle collision kinematics instead of scalar wear rates. Specifically, we project and compress 23 Eulerian boundary cross-moments (𝔼[V_p^u sin^vα_p cos^wα_p]) across the pipe surface. Using 372 high-fidelity CFD-DPM cases of 90^∘ elbows over three bend ratios (R/D ∈ {1.5, 2.0, 5.0}), five Reynolds numbers, five density ratios, and six particle diameters in the inertial regime (St > 1), we evaluate a hybrid compression scheme. Linear Proper Orthogonal Decomposition (POD) and Mode-1 tensor unfolding SVD are combined with block-wise Convolutional Autoencoders (CNN-AE) to handle both broad convective transport and localized impact craters. An anisotropic Gaussian Process Regression (GPR) surrogate maps four dimensionless Π-groups to the compressed latent space, evaluating full 2D wear topographies in roughly 2 ms (R^2 > 0.99 on primary kinematic fields). Because kinematics are decoupled from material damage laws, the resulting surrogate evaluates multiple empirical models post-hoc exactly matching Finnie and closely approximating Oka, McLaury, and Arabnejad without retraining.
ER -