Accepted Abstracts

Poclain Hydraulic is an independent industrial group specialized in the design, the manufacturing and the marketing of efficient transmissions (hydrostatic, electrohydraulic, full electric) headquartered in France. Poclain Hydraulics strives to offer a complete solution for most-demanding applications. An increase of market needs and demands require hydraulic components and systems to be more efficient, more reliable, and last but not least, adopted to customer environment.

Positive displacement machines are mechanical devices that move fluid (liquid or gas) by “trapping” a fixed amount of it and then displacing it into a discharge line. These machines work by physically capturing a certain volume of fluid and transfer it from upstream to downstream area.
Simulation of such machines still requires highly skilled engineers, use of advanced simulation tools and advanced simulation approach … despite the availability of modern simulation tools with well suited graphical user interface (GUI) and numerical techniques.

The paper presents recent activities and progress on simulation of positive displacement machines – the axial piston pump and the radial hydraulic motor, in particular. Despite that these machines have been designed and produced by Poclain for decades, there are still (design) features and phenomena not being investigated in detail or never being simulated.

The simulation advancements mainly refer to the application of complicated kinematic motion, fluid properties, physics to consider as well as mesh and numerical algorithm techniques. In this paper, the focus is given on modelling of advanced fluid properties (e.g. compressibility, cavitation etc.). Fluid with such advanced properties is used in fluid flow simulation.

Numerical approach has been performed by means of computational fluid dynamic (CFD). The cavitation detection has been possible by introducing (implementation) of existing “full cavitation model”, developed by Singhal et al. Results for axial piston pump shows good agreement with experimental investigation for piston chamber pressure and tilting moment on a swash plate.

In the pursuit of sustainable and cost-effective engineering solutions, the substitution of metal components with high-performance plastics has become a strategic objective across various industries. This presentation introduces a comprehensive case study from X-Plast showcasing the successful redesign of a highly loaded railway component originally manufactured from metal. The project, developed in collaboration with Knorr-Bremse Rail Systems, aimed to reduce weight and cost while maintaining structural integrity and ensuring a 30-year operational lifespan under extreme environmental conditions.

The transformation from a simple metal plate to a robust plastic part was achieved through X-Plast’s integrated development process, which combines simulation-driven design, rapid prototyping, and advanced manufacturing optimization. The workflow incorporated state-of-the-art software tools from Altair (Inspire, Simsolid, Simlab, OptiStruct), Autodesk (Moldflow Insight), and PART Engineering (Converse, S-Life Plastics), enabling a seamless transition from isotropic to anisotropic analysis. Key steps included topology optimization, nonlinear contact modeling, and the mapping of injection molding results—such as fiber orientation, weld lines, and residual stresses—into the finite element model.

This function-driven approach allowed for rapid iteration and validation of design variants, significantly shortening development cycles. The final design not only met all mechanical and thermal requirements but also achieved a 35% reduction in production costs and a 25% decrease in component weight. The use of anisotropic material models and lifetime prediction tools ensured reliable performance over decades of service.
Beyond the technical achievements, the project demonstrates the strategic advantage of simulation-supported development in the transportation sector. By integrating process simulation with structural analysis, X-Plast was able to optimize manufacturability and mechanical behavior simultaneously.

Attendees will gain insights into the practical implementation of coupled simulation workflows, including how to bridge injection molding data with finite element analysis, and how to leverage anisotropic modeling for fatigue and lifetime prediction. The case study exemplifies how collaborative engineering, simulation expertise, and tailored software integration can unlock new potentials in plastic part development for demanding applications.

The experimental evaluation of the fitting parameters of composite multiaxial 1st-ply-failure models is a difficult task for several reasons such as lack of standard mechanical experiments and processing techniques, nonlinearity in stress-strain response at failure state, no guidance available on the selection of most informative experiments, contradicting theoretical failure models as well as uncertainty in strength. The presented study offers a holistic view and solution to the issues above. It comprises a nonlinear Finite Element Analysis (FEA) based failure stress state evaluation, then the appropriateness of the available 1st-ply-failure models is checked using local sensitivity, Fisher Information Matrix (FIM) theory and robustness analysis. As next, the most informative subset of experimental failure stress states is selected using nonlinear Design of Experiments (DoE) technique for a more accurate nominal failure parameter fit. Finally, a statistical Markov-Chain-Monte-Carlo (MCMC) approach is implemented to evaluate the expected distribution of all in-plane strength components, thus, the likely range of occurrence of them.

The entire approach is demonstrated using experimental data of an IM7/8552 carbon fiber (CF) and epoxy system. Beside the standard tension and compression, as well as shear experiments, the dataset was extended with off-axis tension and compression test data of unidirectional (UD) samples and with special multiaxial tension/compression – torsion measurements performed on composite tubes. It was crucial to find a bespoke composite layup of the tube specimens at the grip area to withstand load concentrations and to keep failure location in the gauge section. In addition, in compression dominated experiments the potential buckling of test pieces was also mandated to avoid. FEA was used to find the best tube designs that satisfied all expectations. These simulations are also briefly presented.

The failure stress state of the samples was estimated using the failure load information and FEA of each specimen experiment. In order to stay as accurate as possible, nonlinear orthotropic material constitutive model was defined for the FEA representation of the specimens. The parameters of the nonlinear orthotropic material card were evaluated with prior constitutive material model fitting to standard test data.
The failure model fitting entire approach is implemented in an automated system that is also briefly introduced. Based on the outcome of this study it was concluded that without the information from special multiaxial experiments (such as tube torsion) the failure envelope cannot be fitted with adequate accuracy. This can lead to the inappropriate strength prediction in multiaxial states – typically when normal stress is combined with shear – and may also result in overestimation of the maximum shear stress.

Poclain Hydraulic is an independent industrial group specialized in the design, the manufacturing and the marketing of efficient transmissions (hydrostatic, electrohydraulic, full electric) headquartered in France. Poclain Hydraulics strives to offer a complete solution for most-demanding applications. An increase of market needs and demands require hydraulic components and systems to be more efficient, more reliable, and last but not least, adopted to customer environment.

Innovation, the company's spearhead, is deeply rooted in the group's core values. Simulation-based design (upfront simulation) is well-placed across design teams which mean that each design engineer is able to perform basic simulations by himself. The path to such a simulation democratization is a long-term process, where one of the main challenge it to change people mindset.

Upfront simulation is a term to describe simulation activities being performed early in the design stage. So before physical prototypes are made or full systems being built. The goal is to identify potential issues, optimize designs and predict performance as early as possible, when changes are cheaper and easier to make.

Simulation democratization is understood as to make simulation tools accessible to non-expert users within enterprise, organization or community. It means removing the barriers that restrict product managers, designers or even students to run simulations and make data-driven decisions without needing to be simulation experts.

The paper presents current status regarding upfront simulations and their democratization within Poclain. The practical cases demonstrate the abilities brought through the upfront (or frontloading) simulations and what activities have been made to democratize (deploy) them among the group of designers and design engineers.
The power of upfront simulations is also demonstrated on real case example, performed in real time (in-situ) at the conference. It emphasizes its ease of use, the importance to understand the product behaviour as well as the gain in time to market (due to short time from pre- to post-processing).

The growth of computer-aided engineering (CAE) tools and simulation data offers new opportunities for engineering teams to develop new products faster. However, challenges persist due to fragmented workflows, siloed simulation data management, and inefficient manual metadata processes. Engineers often spend up to most of their time managing data [1] instead of making critical design decisions, impacting time-to-market, operational costs, and the ability to deliver optimal solutions.

Despite its many advantages, the use of CAE in the product development process presents several challenges. One major hurdle is the high computational power and specialized software required for advanced simulations. Additionally, accurately modeling real-world conditions in a virtual environment is complex and may not always capture all the nuances of physical behavior, leading to discrepancies between simulation results and real-world performance. CAE also relies heavily on accurate material data and boundary conditions; any errors or assumptions in these inputs can lead to inaccurate predictions. Lastly, while CAE speeds up development by reducing the need for physical prototypes, it may still require validation through physical testing, which can introduce delays and costs. These challenges highlight the need for continuous improvement in CAE technologies and the expertise of engineers to fully harness its potential.

In this abstract, we will talk about how a Graph Neural Network (GNN) architecture is deployed for modeling the complex behavior of bipolar plates of PEM (Proton Exchange Membrane) fuel cells. Modeling fuel cells involves complex FEA and CFD methods. Geometry preparation for the FEA process is human intensive and solving the FEA simulation takes a minimum of 48 hours on 100s of CPUs. The deformed geometry from the FEA simulation is processed into a CFD model for the flow prediction. Using a surrogate model approach we will demonstrate how we can predict the structural deformation of the geometry starting from the CAD model. Inferencing on new designs with the surrogate model reduces design validation time from 48 hours with traditional methods to just seconds in real-time.

A framework to streamline multidisciplinary simulation workflows by integrating digital thread concepts and AI-driven methodologies is discussed. By unifying historical modeling and simulation data, automating metadata capture, and leveraging AI for optimization, this framework approach significantly enhances collaboration, decision-making, and productivity.

The increasing demand for lightweight yet mechanically robust components in the aerospace and automotive industries requires an integrated understanding of both structural and thermal behaviors under realistic service conditions. This study presents a coupled structural–thermal simulation framework developed to analyze and optimize the thermo-mechanical performance of lightweight materials, with a focus on aluminum alloy and composite structures.

Using a multiphysics finite element and computational fluid–thermal (FEM/CFD) approach, the model couples transient heat transfer with mechanical stress analysis to capture temperature-dependent deformation, thermal expansion, and stress concentration. The simulation workflow includes temperature-dependent material properties, thermal boundary conditions reflecting convection and heat flux, and structural loads representative of real operating scenarios. Parametric studies are performed to evaluate the influence of geometric design, thickness variation, and boundary conditions on structural integrity and thermal stability.

Results show that coupling thermal and structural fields provides a more accurate prediction of component behavior compared to single-physics models. Optimization based on response-surface methodology identifies design configurations that minimize both thermal gradients and maximum von Mises stress while reducing overall weight.

The proposed coupled framework demonstrates an effective computational strategy for improving the reliability and performance of lightweight components in thermally dynamic environments. The methodology can be extended to advanced applications such as electric vehicle battery housings, aerospace panels, and heat-sensitive mechanical assemblies, where precise thermal-structural interaction modeling is essential for product development.

The development of smart tires is reshaping the future of mobility, especially in the context of electric and autonomous vehicles. This paper presents a simulation-driven approach to tire engineering that integrates advanced 3D FEA modeling with sensor signal analysis and big data workflows. Using Abaqus Explicit, we demonstrate how legacy modeling limitations can be overcome to enable system-level design and real-time performance evaluation. Two sensor mounting scenarios are explored, along with the concept of a Virtual Intelligent Tire Belt equipped with multiple acceleration sensors. These configurations generate extensive datasets across varied load cases, highlighting the need for automated post-processing and efficient data management. Our workflow supports the analysis of static and dynamic tire load cases and sensor responses under complex loading conditions. The proposed methodology enables predictive analytics, supports ADAS algorithm development, and aligns with CASE mobility requirements. By combining simulation, automation, and data science, this work offers a scalable solution for smart tire R&D and contributes to the broader goal of engineering innovation in the automotive sector.

As simulation becomes increasingly central to engineering decision-making, the need for structured Simulation Data Management (SDM) is growing. Yet many organizations (especially those early in their SDM journey) still rely on local storage or shared drives, lacking traceability, version control, and collaboration tools. This presentation introduces a practical SDM solution built on Windchill and Navigate, designed to meet the needs of CAE engineers while leveraging existing PLM infrastructure.

The increasing need for reliability and efficiency in wind assets highlights the importance of turbine-level monitoring frameworks capable of detecting anomalies, evaluating degradation, and improving power output prediction. This paper presents the development of a Digital Twin (DT) specifically for one wind turbine located in the Bogdanci Wind Farm in North Macedonia. Unlike farm-scale DT platforms, this work focuses on a single-unit replicative model, enabling high-resolution analysis of turbine behavior, operational deviations, and localized performance optimization strategies. The proposed DT follows a hybrid modelling architecture consisting of four core layers:

i) Physics-based analytical turbine model, capturing the aerodynamic performance through power coefficient curves as a function of wind speed, blade pitch, and rotor dynamics. The model acts as a simulation baseline that can generate theoretical power output for variable inflow conditions and operational control settings.

ii) Machine-learning prediction module, trained exclusively on SCADA measurements from the target turbine. Early-stage Python workflows include filtering, SCADA time-series alignment, derivation of turbulence intensity and directional wind distribution, and regression-based power prediction. Initial results show reduced deviation from measured output, especially during partial load and turbulent inflow conditions, compared to classical power curve estimation.

iii) Diagnostics through residual comparison, where deviations between simulated and ML-predicted output form a basis for anomaly tracking. Detected patterns correspond to potential operational issues such as yaw offset, soiling-related power loss, and thermal derating. A lightweight alerting mechanism is under development to flag performance drops in near-real-time.

iv) Visualization and reporting component, currently implemented as a prototype dashboard for plotting power curve displacement, SCADA variability metrics, turbine efficiency over time, and synthetic-vs-measured energy yield.

Preliminary findings confirm that single-turbine digital twinning enables detailed insight into unit-specific behavior otherwise masked in aggregated wind farm data. The hybrid analytical–ML model improves performance estimation accuracy, opens a pathway toward turbine-level predictive maintenance, and establishes a methodological basis for scaling toward multi-turbine or full-farm digital twins. Future work includes refinement of anomaly-trigger thresholds, integration of short-term forecasting models, and extension toward multi-asset hybrid platforms combining wind, solar, and storage.

The mechanical characterization and constitutive model selection of elastomers is a significant challenge in engineering simulations, since those materials can exhibit not only nonlinear finite-strain elastic behavior, but also time-dependent (viscoelastic) properties in combination with substantial temperature dependency. A common approach for modelling such behaviour is to combine a visco-hyperelastic material model with temperature-dependent parameters. During the parameter identification process, the visco-hyperelastic parameters are fitted to measurement data at various temperature levels, and for intermediate temperature levels, the parameters are usually interpolated. This method can be efficient for simpler hyperelastic models (e.g., Neo-Hookean, Arruda-Boyce); however, for models with a higher number of parameters (e.g., higher-order Ogden, Polynomial), intermediate temperature levels may result in physically non-admissible stress solutions.

Therefore, it is aimed to perform temperature-dependent model fitting and enable prediction at intermediate temperatures where no actual measurements have been conducted. To achieve this, virtual measurements are generated and included in the fitting. These virtual measurements are calculated by linear interpolation between measured curves at two adjacent temperatures.

To perform temperature-dependent fitting, it is necessary to select a reference temperature at which the fitting is first performed. After fitting, the parameter values obtained at this temperature are used to define the constraints at other temperatures. In the proposed method, the model parameters can depend on temperature in three ways: constant, linearly constrained, or unconstrained. If a parameter is constant, its value is equal to the parameter value at the reference temperature for all temperatures. For a linear constraint, it is identified by constraining it to have linear temperature dependence. The proposed fitting algorithm determines the value of the slope that provides the best fitting parameter values across all temperatures. If a parameter has no constraint, the result at the reference temperature is used as an initial condition for fitting the model to the measurements at the other temperatures. For intermediate temperatures between two measured temperatures (where only virtual measurements are available), the limits of the parameter value are prescribed so that its value does not change significantly between the two adjacent temperatures.

In this contribution, a detailed methodology for parameter identification of temperature-dependent visco-hyperelastic materials is provided. The proposed method and the effect of the temperature-dependent material constraints are demonstrated via a case study based on temperature-dependent measurements performed on an elastomer material with various load cases and temperature levels. Finally, the efficiency of the proposed method is also compared with the classical method.

Cerebral aneurysms are estimated to affect approximately 2–5% of the general population [1, 2]. While these lesions often remain asymptomatic, the rupture of an intracranial aneurysm leads to subarachnoid hemorrhage (SAH), a condition associated with high morbidity and mortality. Consequently, discriminating between stable aneurysms and those at imminent risk of rupture remains a critical challenge in neurosurgery. Because standard clinical metrics often fail to capture the complex biomechanical environment of the vessel wall, advanced Fluid-Structure Interaction (FSI) simulations are increasingly utilized to predict wall integrity and stability non-invasively.
The initiation and progression of cerebral aneurysms are driven by complex hemodynamic forces, particularly the interplay between Wall Shear Stress (WSS) and the Oscillatory Shear Index (OSI). While earlier literature debated the dominance of high versus low WSS theories in aneurysm pathogenesis [3], recent FSI analyses suggest a reciprocal regulation between these two metrics. It is observed that while high WSS often triggers the initial formation, the subsequent growth phase is frequently characterized by a transition to low WSS and high OSI environments [4]. This shift suggests that oscillatory shear is not merely a flow artifact but a primary driver of vascular remodeling, making it a critical parameter for evaluating aneurysm maturity.
To leverage this hemodynamic insight for clinical assessment, a validated FSI workflow was employed to analyze patient-specific geometries. A novel metric, the "Aneurysm-OSI-Band" (AOB), was identified to map wall composition non-invasively [5]. Defined by a distinct band of high OSI (>0.1) at the aneurysm neck that transitions homogeneously to lower values in the dome, the AOB was validated against intraoperative video analysis. It was demonstrated that this hemodynamic pattern correlates statistically with wall thickness: high-OSI regions correspond to thick, stable tissue, whereas low-OSI regions identify thinner, more vulnerable areas.
The aim of this work is to show the potential of the AOB as a predictive marker for rupture risk evaluating a retrospective cohort of 100 Middle Cerebral Artery (MCA) aneurysms [6]. The presence of an AOB is assessed across three clinical subgroups: ruptured, unruptured-treated, and unruptured-stable. The analysis reveals a statistically significant correlation between the presence of an AOB and aneurysm stability. An AOB was absent in 100% of the ruptured aneurysms (n=23) and was present in only 7% of the unruptured-treated cases (n=57). In stark contrast, 65% of the stable, unruptured aneurysms (n=20) exhibited this protective feature. Statistical evaluation using a Mann-Whitney U-test confirmed a significant difference (p < 0.0001) between stable aneurysms and the ruptured/treated groups.
These findings suggest that the Aneurysm-OSI-Band may serve as a robust indicator of structural stability. The absence of an AOB implies a heterogeneous, potentially thin-walled structure with a significantly elevated risk of rupture. By integrating FSI-derived hemodynamic metrics into preoperative planning, a more precise, non-invasive risk stratification can be achieved. Future work will focus on multicenter validation to confirm these thresholds and the integration of these metrics into automated clinical software to support decision-making regarding conservative management versus surgical intervention.

References
[1] Dhar S et al, doi: 10.1227/01.NEU.0000316847.64140.81
[2] Vlak MH et al, doi: 10.1016/S1474-4422(11)70109-0.
[3] Meng H et al, doi: 10.3174/ajnr.A3558.
[4] Nagy J et al, doi: 10.3390/brainsci14100977.
[5] Nagy J et al, doi: 10.59972/bkzfvrqx
[6] Nagy J et al, doi: 10.1038/s41598-024-85066-9.