A multi-fidelity data-driven constitutive modeling framework via physics encoding and transfer learning: application to geotechnical engineering problems

T. Wichtmann, T. Triantafyllidis: Behaviour of granular soils under environmentally induced cyclic loads. - D. Muir Wood: Constitutive modelling. - C. di Prisco: Creep versus transient loading effects in geotechnical problems. - M. Pastor et al.: Mathematical models for transient, dynamic and cyclic problems in geotechnical engineering. - M. Pastor: Discretization techniques for transient, dynamics and cyclic problems in geotechnical engineering: first order hyperbolic partial diffential equations. - M. Pastor et l.: Discretization techniques for transient, dynamic and cyclic problems in geotechnical engineering: second order equation. - C. di Prisco: Cyclic mechanical response of rigid bodies interacting with sand strata. - D. Muir Wood: Macroelement modelling. - M. F. Randolph: Offshore design approaches and model tests for sub-failure cyclic loading of foundations. - M.F. Randolph: Cyclic interface shearing in sand and cemented solis and application to axial response of piles. - M. F. Randolph: Evaluation of the remoulded shear strength of offshore clays and application to pipline-soil and riser-soil interaction.

The integration of numerical modeling and physical measurements through inverse analysis in geotechnical engineering

Data-driven multiscale modelling of granular materials via knowledge transfer and sharing

Green infrastructures are decentralized semi-natural solutions for managing stormwater runoff. To assess their effectiveness in urban drainage, many process-based hydrological models have been developed. However, these models sometimes fail to deliver satisfactory results due to inadequate representations of the involved hydrological processes or due to limited field measurements for model setup. Data-driven modelling simulates directly the connections between the state variables of the system (such as the input variable and the output variable), and thus reduces the need for hydrological process characterization. Yet the usefulness of this modelling approach in green infrastructure related studies has not been fully explored. To demonstrate its effectiveness, two bioretention systems in the U.S. are studied. The designs of these two systems are unique, making them difficult to be modelled directly using the standard process-based models. In one site, the observations are also censored, i.e., only the overflow/no overflow binary outcomes are being monitored, with which the model calibration is inhibited. We showed that using the state-of-the-art data-driven modelling framework (e.g., CARET and MXNET), the occurrence or the outflow discharge rate of the bioretention systems at a given time can be accurately simulated only using the rainfall, evapotranspiration and the inflow time series. We conclude that data-driven modelling can be useful for simulating the hydrologic performance of green infrastructures, especially when the existing process-based models are inadequate in representing the design variabilities or the data for model setup is limited.

Accurate identification of process parameters from in-mold sensing signals is essential for advancing intelligent and adaptive injection molding systems. This study proposes a data-driven inverse modeling framework that infers key process parameters directly from cavity pressure curves. A stage-wise Autoencoder (AE) is employed to perform nonlinear feature extraction by segmenting the pressure curve into filling, packing, and cooling stages, thereby capturing process-dominant dynamics. The encoded latent features are subsequently mapped to injection speed, packing pressure, nozzle temperature, and mold temperature using a multilayer perceptron (MLP). Eighty-one full-factorial parameter combinations were experimentally conducted, generating 810 molding cycles on an industrial injection molding machine. An independent validation dataset was collected from separate production runs to evaluate generalization capability. The proposed AE–MLP framework achieved high predictive accuracy on unseen data, with R² values of 0.9979 for injection speed, 0.9926 for packing pressure, 0.9571 for mold temperature, and 0.9296 for nozzle temperature. The corresponding RMSE values remained below 1.6 across all parameters. Comparative analysis against PCA–MLP and direct MLP models demonstrates that nonlinear manifold representation significantly improves inverse regression robustness, particularly for thermally coupled parameters. Furthermore, the stability of inverse modeling is analyzed from the perspective of parameter identifiability and many-to-one mapping characteristics inherent in injection molding processes. Results indicate that stage-wise nonlinear feature encoding enhances separability in the latent space and mitigates ambiguity in parameter inference. The proposed framework provides a robust and physically interpretable approach for real-time process parameter estimation, contributing to the development of adaptive and self-optimizing manufacturing systems. Infers molding parameters directly from multi-sensor pressure features. AE–MLP framework learns nonlinear relations between pressure and parameters. Parameter recommendation enables adaptive process condition adjustment. 810-cycle full-factorial experiments confirm accuracy and generalization. Supports smart manufacturing through real-time parameter tuning.

Accurate wind speed forecasting plays a central role in the integration of wind energy into modern power systems. However, in many regions, the deployment of data-driven models is limited by the scarcity of historical wind measurements. This study investigates the potential of transfer learning (TL) to address this issue by reusing models trained on data-rich locations to forecast wind speed in under-instrumented sites. Using a dataset from 70 meteorological stations across Spain, spanning diverse climatic conditions, we compared TL with conventional direct learning (DL) using Extreme Learning Machine (ELM) and Autoregressive (AR) models. Forecasts were performed over horizons ranging from 30 min to 6 hr and evaluated using normalized Root Mean Squared Error (nRMSE) and normalized Mean Absolute Error (nMAE). Across 4,830 TL experiments, results show that TL achieves forecast accuracy close to that of DL, with average nRMSE ranging from 0.297 (TL) to 0.292 (DL) for 30-min horizons, and from 0.674 to 0.640 for 360-min forecasts. These findings confirm that TL is a robust and scalable approach for wind speed forecasting in data-scarce environments. The methodology opens promising avenues for the deployment of forecasting tools in regions where traditional data-driven models remain inapplicable due to limited local measurements.

A novel image-based transfer learning framework for cross-domain HVAC fault diagnosis: From multi-source data integration to knowledge sharing strategies

STEAM: A scalable data-driven surrogate modeling framework for part-scale scanwise thermal process simulation of laser powder bed fusion

Optimization techniques have drawn much attention for solving geotechnical engineering problems in recent years. Particle swarm optimization (PSO) is one of the most widely used population-based optimizers with a wide range of applications. In this paper, we first provide a detailed review of applications of PSO on different geotechnical problems. Then, we present a comprehensive computational study using several variants of PSO to solve three specific geotechnical engineering benchmark problems: the retaining wall, shallow footing, and slope stability. Through the computational study, we aim to better understand the algorithm behavior, in particular on how to balance exploratory and exploitative mechanisms in these PSO variants. Experimental results show that, although there is no universal strategy to enhance the performance of PSO for all the problems tackled, accuracies for most of the PSO variants are significantly higher compared to the original PSO in a majority of cases.

SummaryTo solve large deformation geotechnical problems, a novel strain‐smoothed particle finite element method (SPFEM) is proposed that incorporates a simple and effective edge‐based strain smoothing method within the framework of original PFEM. Compared with the original PFEM, the proposed novel SPFEM can solve the volumetric locking problem like previously developed node‐based smoothed PFEM when lower‐order triangular element is used. Compared with the node‐based smoothed PFEM known as “overly soft” or underestimation property, the proposed SPFEM offers super‐convergent and very accurate solutions due to the implementation of edge‐based strain smoothing method. To guarantee the computational stability, the proposed SPFEM uses an explicit time integration scheme and adopts an adaptive updating time step. Performance of the proposed SPFEM for geotechnical problems is first examined by four benchmark numerical examples: (a) bar vibrations, (b) large settlement of strip footing, (c) collapse of aluminium bars column, and (d) failure of a homogeneous soil slope. Finally, the progressive failure of slope of sensitive clay is simulated using the proposed SPFEM to show its outstanding performance in solving large deformation geotechnical problems. All results demonstrate that the novel SPFEM is a powerful and easily extensible numerical method for analysing large deformation problems in geotechnical engineering.

Trivariate analysis of soil ranking-correlated characteristics and its application to probabilistic stability assessments in geotechnical engineering problems

Accurately predicting building heating and cooling loads is crucial for optimizing HVAC systems and enhancing energy efficiency. However, data-driven models often face overfitting issues due to scarce training data, a common challenge for new constructions or under data privacy constraints. Transfer learning (TL) offers a solution, but its effectiveness heavily depends on selecting an appropriate source domain through effective similarity measurement. This study systematically evaluates the performance of 20 prevalent similarity metrics in TL for building heating load forecasting to identify the most robust metrics for mitigating data scarcity. Experiments were conducted on data from 500 buildings, with seven distinct low-data target scenarios established for a single target building. The Relative Error Gap (REG) was employed to assess the efficacy of transfer learning facilitated by each metric. The results demonstrate that distance-based metrics, particularly Euclidean, normalized Euclidean, and Manhattan distances, consistently yielded lower REG values and higher stability across scenarios. In contrast, probabilistic measures such as the Bhattacharyya coefficient and Bray–Curtis similarity exhibited poorer and less stable performance. This research provides a validated guideline for selecting similarity metrics in TL applications for building energy forecasting.

The industrial sector is nowadays experiencing a digital transformation motivated by the Industry 4.0 paradigm. Concepts such as data-driven models, Artificial Neural Networks (ANNs), and Transfer Learning (TL) are part of the current vocabulary in the industrial management and control topics. For that reason, in this paper the application of TL techniques is proposed to derive new ANN-based control structures from pre-existing ones. Notice that if an ANN-based controller is transferred into a new industrial environment, its appropriate behaviour must be ensured, and what it is more important, this must be known a priori. Nevertheless, TL techniques do not always ensure this. That is why the Transfer Suitability Metric (TSM) is proposed here. Determining the similarity among environments, this metric tells if the controller can be transferred, transferred with certain limitations, or if it cannot be transferred at all. Here, the metric is applied over a Wastewater Treatment Plant (WWTP). The objective is to derive the control structure of one control loop, let us say the Dissolved Oxygen (DO), and then transfer it into another basic control loop in a WWTP, the Nitrate–nitrogen (NO), and vice-versa. Results show that with the help of the TSM, an improvement around a 68.54% and 80.53% in the Integrated Absolute Error (IAE) and the Integrated Squared Error (ISE) is obtained in the NO management, respectively. Moreover, a simplification and speed-up of the controller design process is achieved.

Comprehensive transferability assessment of short-term cross-building-energy prediction using deep adversarial network transfer learning

Data-driven intelligent modeling framework for the steam cracking process