Umeå University's logo

Piping is one of the most common and hazardous issue in dyke and dam engineering, posing challenges for dyke and dam stability and risk assessments. In this study, an interpretable ensemble learning prediction model of dyke and dam piping was proposed based on the Synthetic Minority Over-sampling Technique (SMOTE) method and Ensemble Learning (EL) algorithm with a dataset collected from Yangtze River. Initially, the piping dataset was visualized using the violin diagram, and the SMOTE method was adopted to augment the imbalanced dataset. Then, t-distributed Stochastic Neighbor Embedding (t-SEN) method and Pearson correlation coefficient were used to consider the similarity between the newly generated samples and the original samples, which verify the effectiveness of the data augmentation. Subsequently, based on the augmented dataset, six EL algorithms were employed to establish the regression prediction model of piping. Through comprehensive comparison, the SMOTE-Categorical Boosting (SMOTE-CatBoost) model exhibits superior prediction accuracy and lower calculation cost, with a goodness of fit (R2) of 0.9886 and a Root Mean Square Error (RMSE) of 0.05334, making it the ideal prediction model for dyke and dam piping. Additionally, an Explainable Artificial Intelligence (XAI) model of piping was developed, and it was found that the thickness of overburden thickness of weak permeable layer (H), void ratio (e), water level height difference (Δh), and compression coefficient (av) are the four primary influencing factors of piping. The research offers valuable reference for the advance monitoring of dyke and dam piping risk, and contributes to the sustainable maintenance of dyke and dam engineering structures.

Ultraviolet-cured glass fiber reinforced polymer (UV-GFRP) composites are widely used in cured-in-place pipe (CIPP) repair technology for buried pipelines. The bending strength is the key indicator for assessing repair quality, which is affected by multiple factors but lacks effective prediction methods yet. In this paper, a prediction model for the bending strength of UV-GFRP composites based on the archimedes optimization algorithm (AOA) combined with the extreme gradient boosting (XGBoost) algorithm is proposed, incorporating material structure design and curing parameters. Through hyperparameter optimization, robustness analysis, and sensitivity analysis, the model's performance and reliability are thoroughly evaluated. The results show that the AOA-XGBoost model achieves highly accurate prediction, with an R² of 0.906 on the test set, outperforming the backpropagation neural network optimized by genetic algorithm (GA-BPNN), support vector regression optimized by particle swarm optimization (PSO-SVR), random forest regression (RFR), gradient boosting decision tree (GBDT), and XGBoost. Notably, the model maintains stable predictions even under noise conditions of up to 10%. Sensitivity analysis reveals that fiber volume fraction (+0.338), glass fiber architecture (+0.205), and density (+0.178) have the most significant effect on the bending strength of UV-GFRP composites, which can be optimized to enhance material properties. Although curing parameters have a relatively smaller effect, careful adjustment is essential to prevent over-polymerization and degradation of material properties.

Polyurethane (PU) as a popular polymer is widely recognized for its exceptional thermal insulation properties, making it a critical material for applications requiring effective heat transfer management. Integrating Phase Change Materials (PCMs) into PU (PU-PCMs) has emerged as a highly effective strategy for enhancing building envelope performance, ensuring greater indoor thermal stability, and mitigating temperature fluctuations. This study introduces an Enhanced Multi-Scale Modeling approach to investigate the thermal conductivity and energy efficiency of PU-PCMs, with a particular focus on their comparative performance in subarctic and tropical climates. By combining Molecular Dynamics (MD) simulations with Finite Element Methods (FEM), the model seamlessly bridges molecular-scale interactions with macroscopic thermal behavior. Using a Representative Volume Element (RVE)-FEM framework, microscopic properties are translated into engineering-scale parameters, enabling accurate and efficient predictions of material performance across diverse climatic conditions. The findings demonstrate the capability of PU-PCMs to significantly enhance energy efficiency and indoor thermal comfort. In a case study of a single-family house, PU-PCMs achieved a 7.376% reduction in energy consumption and increased comfort hours under Stockholm's subarctic climate. Moreover, the adaptability of PU-PCMs was validated in both tropical and subarctic environments, consistently stabilizing indoor temperatures, reducing HVAC energy demands, and improving occupant comfort. These results underscore the potential of PU-PCMs as a passive thermal management solution, advancing sustainable building practices. The proposed multi-scale model offers a computationally efficient and precise tool for optimizing material design in energy-sensitive applications, reinforcing the versatility and significance of PU-PCMs across a range of climatic conditions.

Nanoreinforced polymer composites have been widely investigated and developed due to their excellent physical and chemical properties, and they have been broadly applied in aerospace engineering, microelectronic packaging, and electronic and semiconductor devices. In recent years, numerous studies have focused on quantifying the influence of nanoscale fillers on the performance of these composites. However, deterministic models often overlook material uncertainties, and stochastic multiscale modeling is computationally expensive. With the advancement of high-performance computing, machine learning—well-known for its efficient modeling capabilities—has become increasingly popular for addressing these uncertainties.

Based on this motivation, we developed a multiscale and multi-tier framework that integrates theoretical modeling with experimental applications. Functional nanocomposites are designed from the micro- to macro-scale according to engineering requirements, followed by laboratory testing of their properties. The tested materials are then deployed in real environments for field measurements and application validation. The results demonstrate that this multiscale and multi-tier framework is capable of designing optimal materials tailored to engineering needs while enabling a seamless transition from theoretical simulation to experimental implementation for polymer nanocomposites.

Our study highlights an effective pathway for the design and development of nanoreinforced polymer composites through a multiscale and multi-tier strategy, allowing them to better meet practical engineering demands. This approach provides new ideas and tools for the future design and development of intelligent advanced materials and is expected to further promote the application and evolution of functional nanocomposites across various fields.

The advancement of artificial intelligence (AI) in material design and engineering has led to significant improvements in predictive modeling of material properties. However, the lack of interpretability in machine learning (ML)-based material informatics presents a major barrier to its practical adoption. This study proposes a novel quantitative computational framework that integrates ML models with explainable artificial intelligence (XAI) techniques to enhance both predictive accuracy and interpretability in material property prediction. The framework systematically incorporates a structured pipeline, including data processing, feature selection, model training, performance evaluation, explainability analysis, and real-world deployment. It is validated through a representative case study on the prediction of high-performance concrete (HPC) compressive strength, utilizing a comparative analysis of ML models such as Random Forest, XGBoost, Support Vector Regression (SVR), and Deep Neural Networks (DNNs). The results demonstrate that XGBoost achieves the highest predictive performance ((Formula presented.)), while SHAP (Shapley Additive Explanations) and LIME (Local Interpretable Model-Agnostic Explanations) provide detailed insights into feature importance and material interactions. Additionally, the deployment of the trained model as a cloud-based Flask-Gunicorn API enables real-time inference, ensuring its scalability and accessibility for industrial and research applications. The proposed framework addresses key limitations of existing ML approaches by integrating advanced explainability techniques, systematically handling nonlinear feature interactions, and providing a scalable deployment strategy. This study contributes to the development of interpretable and deployable AI-driven material informatics, bridging the gap between data-driven predictions and fundamental material science principles.

Graphene-based polymer nanocomposites show great potential for thermal management, but accurately predicting their thermal conductivity remains challenging due to multiscale structural complexity and parameter uncertainty. We propose an innovative approach integrating interpretable stochastic machine learning with multiscale analysis to predict the macroscopic thermal conductivity of graphene-based polymer nanocomposites. Our bottom-up framework addresses uncertainties in meso- and macro-scale input parameters. Using Representative Volume Elements (RVEs) and Finite Element Modeling (FEM), we compute effective thermal conductivity through homogenization. Predictive modeling is powered by the XGBoost regression tree-based algorithm. To elucidate the influence of input parameters on predictions, we employ SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME), providing insights into feature interactions and interpretability. Sensitivity analyses further quantify the impact of design parameters on material properties. This integrated method enhances prediction accuracy, reduces computational costs, and bridges data-driven and physical modeling, offering a scalable solution for designing advanced composite materials for thermal management applications.

This study offers a comprehensive review of the state-of-the-art developments regarding the application of phase change materials (PCMs) into building envelopes and HVAC systems. Incorporating PCMs into building envelopes can significantly enhance thermal storage capacity and reduce the influence of outdoor temperature fluctuations on interior spaces' thermal conditions. Furthermore, combining PCM-enhanced building envelopes with night ventilation can effectively utilize natural cooling resources, thereby improving both the adaptability and efficiency of PCMs in regulating indoor thermal environments and reducing building energy consumption. An alternative approach for energy conservation involves integrating PCMs directly within ventilation and air conditioning systems. This review offers insights and recommendations for future research, highlighting the necessity of further developments in materials science, system optimization, and real-world application studies to maximize the potential of PCMs in the built environment.

We've devised an interpretable stochastic integrated machine learning method for predicting thermal conductivity in Polymeric Graphene-Enhanced Composites (PGECs) across different scales. Our approach, integrating techniques like Regression Trees methods (Gradient Boosting machine), handles uncertain parameters via a bottom-up multi-scale framework using Representative Volume Elements in Finite Element Modeling (RVE-FEM). To understand factors affecting predictions, we use the SHapley Additive exPlanations (SHAP) algorithm. This method aligns well with training data, offering promise for efficiently designing new composite materials for thermal management.

Polyurethane (PU) is an excellent thermal insulator, and incorporating Phase Change Material (PCM) capsules into PU significantly enhances building envelope performance by improving indoor thermal stability and reducing temperature fluctuations. We propose a hierarchical multi-scale model using Physics-Informed Neural Networks (PINNs) to accurately predict and analyze the thermal conductivity of PU-PCM composites at both micro and macro scales. This approach effectively addresses complex inverse problems and multi-scale phenomena, offering insights that optimize material design. A case study further demonstrates the model's potential in improving thermal comfort and reducing energy consumption in buildings. The successful development of this PINNs-based model holds great promise for advancing PU-PCM applications in thermal energy storage and innovative building insulation design.

We introduce a novel approach that combines interpretable quantitative stochastic machine learning with multiscale analysis to predict the macroscopic thermal conductivity of graphene-enhanced polymer nanocomposites. Our method effectively addresses uncertainties in input parameters across meso and macro scales within a bottom-up modeling framework. By integrating Representative Volume Elements (RVE) with traditional Finite Element Modeling (FEM), we calculate the effective thermal conductivity through homogenization. We further enhance predictive modeling by employing the XGBoost regression tree method. To clarify the influence of input variables on model outcomes, we incorporate SHapley Additive exPlanations (SHAP) and Local Interpretable Model-agnostic Explanations (LIME). Additionally, sensitivity analyses are conducted to assess the impact of design parameters on material properties. This comprehensive approach improves both global and local interpretability, clarifying feature interactions in data-driven and physical models. It reduces the reliance on extensive analytical modeling and simulations, enhancing prediction accuracy and significantly lowering computational costs. Our method holds significant promise for the design of new composite materials optimized for thermal management.