Debiao Meng

Hydrogen storage and transport systems use pipeline steel, where hydrogen embrittlement accelerates fatigue crack propagation and reduces service life. Accurate life evaluation is necessary for safe operation. Conventional approaches, such as experimental fatigue testing, finite-element simulations, and empirical models like the Paris law, are limited by high experimental costs, dependence on often-unavailable hydrogen constitutive laws, and deterministic predictions that do not account for uncertainty.This paper offers a physics-informed neural network (PINN) and ensemble learning framework. The method integrates physical and data-driven modeling using Bayesian model averaging (BMA) with a probabilistic neural network (PNN). Validation on API-X70 and X100 steels, BMA-PINN achieves a log mean squared error (MSE) of 0.0538 and [Formula: see text] of 0.9934 for X100, outperforming other models. PINN captures fatigue life stochasticity with 95th-percentile intervals.

Hydrogen embrittlement is a major failure mode of steel structures in hydrogen‐rich environments. Accurate and rapid assessment of material parameters, such as the fracture toughness of steels in hydrogen environments, is crucial for ensuring the safe operation of hydrogen energy infrastructure. In this study, a computational multi‐fidelity modeling framework is proposed to estimate the hydrogen‐assisted fracture toughness of steels within a defined phase‐field parameter space. A finite element model based on the fracture phase field principle is constructed to obtain high‐fidelity data. A low‐fidelity analytical expression that considers hydrogen‐induced fracture energy reduction and stress‐driven accumulation is developed to obtain a large number of low‐fidelity samples. The proposed multi‐fidelity model achieves higher prediction accuracy than popular models designed for small sample sizes, namely, support vector regression, kernel ridge regression, and Gaussian process regression. Notably, the constructed multi‐fidelity model maintains good predictive performance under the investigated loading‐rate cases. Furthermore, the influence of modeling variables on is thoroughly discussed. In summary, the proposed multi‐fidelity modeling framework provides a powerful tool for accurate, high‐throughput hydrogen embrittlement performance assessment.

In practical engineering, mechanical structures are affected by time-dependent uncertainties and by correlated failure modes. This study proposes an enhanced time-variant reliability analysis framework by coupling a mixed Archimedean Copula model with an adaptive Kriging model driven by the maximum expected prediction error (MEPE) learning function. The mixed Copula combines Gumbel, Clayton and Frank components, so that upper-tail, lower-tail and nearly symmetric dependence can be represented in one model. The MEPE function combines Kriging prediction variance and leave-one-out cross-validation error through a dynamic balance factor, which guides early global exploration and later refinement near the most probable point trajectory. For the planar three-bar structure, the mixed-Copula failure probability at t = 7 is 0.4078, close to the direct-MCS benchmark 0.4059, with a relative error of 0.47%; the number of original limit-state evaluations decreases from 1.00 × 10⁶ to 326, corresponding to a 99.97% reduction. For the two-rod parallel structure, the mixed-Copula failure probability at T = 2 is 0.0987, close to the direct-MCS benchmark 0.0992, with a relative error of 0.50%; the number of original limit-state evaluations decreases to 284. The calculated upper- and lower-tail dependence coefficients of the mixed Copula are 0.1345 and 0.2985 for the planar three-bar structure and 0.0003 and 0.3709 for the two-rod parallel structure. These quantitative results show that the proposed framework can describe nonlinear failure dependence more flexibly than a single Copula while retaining high computational efficiency.

Purpose: This study aims to develop a robust notch fatigue life prediction model for hydro turbine components, specifically addressing fatigue failures in welded joints induced by severe stress concentrations at transition fillets. The research focuses on overcoming the inherent limitations of the conventional theory of critical distance (TCD) in the finite life regime and its heavy dependency on empirical fitting parameters. Design/methodology/approach: An enhanced TCD framework is proposed by incorporating the evolutionary characteristics of the baseline S–N curve to formulate a novel expression for the characteristic length L. A stress gradient threshold is introduced to physically define the effective damage zone, and the effective stress is refined using a root-mean-square formulation. Additionally, a path resampling strategy is implemented to optimize finite element data extraction and eliminate numerical errors associated with mesh sensitivity. The model is validated using experimental results from 04Cr13Ni5Mo/ER316 L welded specimens. Findings: Results demonstrate that predicted fatigue lives exhibit strong agreement with experimental observations, significantly outperforming the conventional line method in accuracy and robustness. The proposed model effectively captures the accelerated damage at notch roots without requiring additional empirical parameter calibration. It achieves high predictive precision comparable to extensively calibrated mainstream frameworks. Originality/value: The study establishes a novel mathematical formulation for critical distance derived from S–N curve evolution, enabling the autonomous determination of L and eliminating empirical fitting requirements. This provides a promising and efficient theoretical framework for the structural health assessment of hydroelectric infrastructures.

Purpose: Accurate reliability evaluation of turbine runners under long-term and complex loads is essential, but reliability analysis of turbine runners often suffers from high computational cost, high input dimensionality and low failure probability. This paper aims to develop an accurate and efficient method for reliability analysis of turbine runners. Design/methodology/approach: A support vector regression (SVR)-based enhanced quasi-Monte Carlo simulation (EQMCS) method is proposed for reliability analysis of turbine runners. An efficient scaling formula is also established to improve the robustness of failure probability assessment. The proposed method is validated through a high-dimensional numerical case and a turbine runner case. Findings: The results show that the proposed method can effectively reduce computational cost while maintaining the accuracy of reliability analysis in high-dimensional problems and improve the efficiency and robustness of failure probability evaluation for complex structures. Originality/value: This paper proposes a novel reliability analysis framework that integrates SVR, EQMCS and an efficient scaling strategy. The method can improve the accuracy, efficiency and robustness of reliability assessment for turbine runners.

Purpose: To characterize the nonlinear damage evolution, a new coupled damage term is proposed within the framework of the Palmgren-Miner (P-M) rule, explicitly considering the interaction between the maximum load, stress amplitude ratio, and frequency ratio. Design/methodology/approach: The robustness and accuracy of the proposed model were validated using experimental results from the welded joints and four additional alloy materials. Findings: By effectively mitigating the non-conservative bias and excessive conservatism inherent in existing methods, this study provides a reliable tool for fatigue life prediction of welded structures under CCF loading spectra. Originality/value: By effectively mitigating the non-conservative bias and excessive conservatism inherent in existing methods, this study provides a reliable tool for fatigue life prediction of welded structures under CCF loading spectra.