Hybrid PINN–GPR Method for Probabilistic Prediction of Concrete Creep Behavior

Probabilistic Fatigue Crack Growth Prediction With Improved Particle Filter and Multiple Models

Predicting fatigue crack growth (FCG) is critical for structural safety and reliability, but it is severely challenged by inherent material uncertainties and the limitations of any single physical model. This study proposes a probabilistic FCG prediction method that integrates an improved particle filter (PF) and a Bayesian fusion for multiple FCG models. To alleviate post‐resampling particle impoverishment, a component‐wise modified Metropolis–Hastings (MMH) move algorithm is introduced as a particle rejuvenation step after resampling, which is designed to preserve particle diversity more effectively than the conventional resampling strategies. Furthermore, the Bayesian fusion dynamically weighs multiple FCG models (Paris–Erdogan, Kujawski, Forman, Nasgro), enabling mitigation of single‐model bias. The proposed method was applied on the FCG prediction of a group of compact‐tension (CT) specimens that have undergone experimental testing. Comparisons between the predicted and experimental results demonstrate that the proposed method provides more accurate FCG prediction in most of the CT specimens and the corresponding 95% confidence intervals (CIs) can encompass the measured crack paths.

Explainable ensemble learning framework for bond strength prediction in 3D printed concrete structures

A comprehensive data-driven framework integrating ensemble machine learning models with systematic hyperparameter sensitivity analysis and explainable artificial intelligence techniques is proposed, demonstrating that the XGB model significantly outperforms the other approaches, achieving superior accuracy and robust generalization.

A multi-physics-informed neural network framework for dual prediction of creep residual life and damage fraction in austenitic stainless steels

Accurate prediction of creep degradation across multiple high-temperature alloy grades remains a significant challenge in structural integrity assessment. Classical empirical and mechanistic models are typically restricted to estimating total creep rupture life and are often material-specific, limiting their generalizability. This study presents a physics-informed neural network (PINN) framework tailored for creep life assessment, enabling simultaneous prediction of creep residual life (CRL) and creep damage fraction (CDF) in nine grades of Austenitic Stainless Steels (AusSS). The scientific contributions are threefold. First, a unified dual-output modeling strategy is introduced to concurrently predict residual life and accumulated damage, enabling integrated failure prognosis beyond classical rupture-life models. Second, a multi-physics-informed feature engineering approach is developed by incorporating physically derived descriptors, including the Larson–Miller parameter (LMP), Monkman–Grant parameter (MGP), and stacking fault energy (SFE), thereby embedding creep mechanics directly into the learning process. Third, a physics-constrained loss formulation is implemented to enforce domain-consistent bounds (CRL > 0 and 0 ≤ CDF ≤ 1), ensuring physically admissible and reliable predictions. The framework is trained and validated on 1179 datapoints spanning diverse compositions and service conditions. The proposed PINN outperforms conventional Machine Learning (ML) and standard deep neural network models, achieving test R² values of 0.933 for CRL and 0.936 for CDF. SHAP-based analysis further confirms the dominant role of physics-informed features. The model provides a generalizable and interpretable tool for high-temperature structural health monitoring and predictive maintenance.

A physics knowledge-based surrogate model framework for predicting multiple time-dependent responses of surrounding rock induced by tunnel excavation

A Data-Driven Model for Predicting Casting Solidification Time and Mold Thermal Stress

This study proposes a data-driven surrogate modeling framework for predicting solidification time and mold thermal stress during low-pressure die casting (LPDC) of aluminum alloy wheels. The methodology employed an optimal Latin hypercube design (OLHD) to sample key parameters including cooling channel geometry and process conditions. A sequential simulation methodology combining ProCAST and Abaqus was implemented to generate a comprehensive dataset of solidification times and thermal stress distributions. Based on this dataset, surrogate models were developed using Support Vector Regression, Kriging, and Polynomial Response Surface Methodology, with their hyperparameters automatically tuned through Bayesian Optimization (BO). The optimized models were rigorously evaluated using four statistical metrics: Coefficient of Determination (R2), Mean Squared Error (MSE), Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE). The evaluation results show that the BO–SVR model demonstrated superior prediction accuracy for both output responses and exhibited exceptional nonlinear fitting capability. This work establishes an effective modeling approach for simultaneous quality and efficiency optimization in wheel manufacturing.

Integrating deep neural network with elastoplastic analysis: a hybrid approach for slope stability analysis

Conventional data-driven methods for slope stability analysis often exhibit an over-reliance on data while neglecting underlying physical principles. To address this limitation, this study proposes a physics-informed neural network (PINN) framework that integrates a neural network surrogate with the elastoplastic deformation mechanism of soil slopes. The governing equations incorporating the elastoplastic constitutive model based on the Mohr–Coulomb yield criterion, along with the boundary conditions, are embedded into the training framework of the PINN. The PINN functions as a surrogate model that requires no pre-constructed training dataset and automatically satisfies both the governing equations and the boundary conditions. Consequently, the developed PINN can directly predict the displacement field of a slope and automatically derive the associated stress–strain fields that comply with the deformation mechanism. These outputs are coupled with the multi-initial point sequential quadratic programming (MSQP) algorithm and the slip surface stress analysis (SSSA) method, enabling the efficient identification of the critical slip surface and the calculation of the corresponding factor of safety (FOS). The proposed method is validated through two illustrative examples. Comparisons of the results with those from commercial software confirm the high accuracy of the proposed method in predicting the stress–strain response and the FOS. This study provides a data-driven and physics-informed paradigm for slope stability analysis grounded in clear physical mechanisms.