Agenda

Solder Joint Reliability Workflow Automation for a Ball-Grid Array (BGA) Package

Presented By
Kajal Khan (Ansys)

Authored By
Kajal Khan (Ansys)
Rajesh Meena
Akhil KS
Bharat Deorukhkar (Ansys)

Solder joint reliability is a critical aspect of electronic packaging, and accurate simulation is essential for ensuring product performance and durability. This work presents an automated workflow developed for solder joint reliability simulation of a ball-grid array (BGA) package using multiple Ansys tools. The major challenge for solder joint reliability simulation is the expertise required for multiple tools and the extensive time needed for an end-to-end manual workflow, which is typically in the order of 3-4 weeks. The proposed workflow streamlines the simulation process by automating key steps such as ECAD to MCAD conversion in Ansys Sherlock, geometry preparation in Ansys Discovery, mesh generation, material assignment, boundary condition, setup, and post-processing in Ansys Mechanical. Additionally, temperature from Ansys Icepak can be read into the workflow to define accurate thermal cycling.

An Integrated Workflow for Power Module Development

Presented By
Voon Hon Wong (Siemens Industry Software)

Authored By
Voon Hon Wong (Siemens Industry Software)
Wilfried Wessel
Florian Bauer
Rod Dudzinski (Siemens EDA)
John Parry (Siemens Digital Industry Software)

Enhancing Safety Assessments for CO2 Pipelines with a Validated Fluid-Structure Interaction Model

Presented By
Aida Nonn (Aixima)

Authored By
Aida Nonn (Aixima)
Philipp Marx (Aixima)

Despite global initiatives to reduce CO2 emissions, a significant "decarbonization gap" remains, requiring improved safety standards for CO2 pipelines to support Carbon Capture, Utilization, and Storage (CCUS) technologies. This paper presents a digital twin of a CO2 pipeline developed using a rigorously validated, fully coupled fluid-structure interaction (FSI) model. This model facilitates rapid virtual testing and reliable assessment of fracture propagation, bypassing the inefficiencies of outdated standards that rely on costly full-scale testing.The digital twin is created using a well-defined parameter calibration methodology for the plasticity and fracture of API X52 pipeline material, incorporating high-quality data from small-scale and instrumented Drop-Weight-Tear (DWTT) tests. The fluid dynamics are modeled in an Euler domain using the GERG-2008 Equation of State (EOS), while the soil-pipe interaction is simulated using the Mohr-Coulomb model to account for onshore environmental conditions. Our FSI model, built using the Coupled Euler-Lagrange (CEL) method, has been validated in full-scale burst tests. It provides comprehensive predictions, including fracture arrest, 3D pressure distribution, and detailed deformation analysis. By filling data gaps in current standards, the validated model serves as the basis for new guidelines, reducing the need for full-scale testing and accelerating the implementation of CCUS projects. It also improves the efficiency and sustainability of safety assessments and pipeline designs, promoting a more robust CCUS infrastructure.

Multiphysics Methodology for Detection of Hotspots on PCB using DCIR and Thermal Simulation

Presented By
Priyanka R (TE Connectivity)

Authored By
Priyanka R (TE Connectivity)
Shruthi G. S. (TE Connectivity)

Thermal analysis on PCB is often used to check the distribution of temperature on the PCB board and its components to detect hot spots. ICs and other components on PCB are smaller as compared to how they used to be years ago and operate at faster frequencies and require more power. As more current is passed through these small devices, the voltage drops across these resistive elements converts power into heat, causing increase in temperature and hotspots. This leads to circuit damage, structural integrity, oxidation of PCB components., etc. The simulation for PCB trace heating is a complex problem and can be simplified by coupling the same with DCIR (Direct current Internal resistance) simulation. The power loss results are obtained from DCIR analysis in Ansys SIWave software, which are then imported to the classic ICEPAK software. The power maps are incorporated in the imported CAD files and electric files in ICEPAK. The correlation with physical test result was in the range of 5-15%. This methodology aids in determining the thermal gradient on PCB boards due to trace heating in different layers of PCB. It also helps in identifying hotspots on PCB as per the power concentration. This method can be used wherever there is a requirement to identify thermal profile on PCBs due to heating of current carrying traces.

Implementation of Digital Twin for Cathodic Protection

Presented By
Zhen Qi Chee (Matcor Technology & Services – A Cetim Company)

Authored By
Zhen Qi Chee (Matcor Technology &
Services – A Cetim Company)
Cheng Siong Chin
Hao Chen
Zi Jie Choong
JJ Chong (Newcastle University)
ShiLiang (Johnathan) Tan (Matcor Technology & Services)

Procor, a Cetim in-house numerical program, uses the boundary element method to verify and optimise cathodic protection (CP) systems, providing effective protection for structures and applications while conserving resources. It acts as a quick decision aid and reduces maintenance through reliable evaluation of the lifetime of CP installations. This is achieved using simulations to predict current potential, anode depletion, and the required current for impressed current cathodic protection systems. With the recent growth of the Industrial Internet of Things, there has been an increase in the digitisation of processes to enhance accuracy and productivity. Applying the digital twin to a CP system allows live data to be sent to Procor, enabling real-time simulation of the CP system. The simulation results illustrate the CP potential field gradient on a 3D model (digital twin), with the 3D potential field displayed in different colors based on the potential level, aiding the user in assessing the protection level of the model. Corrective or preventive measures can then be implemented as needed. In light of these developments, there is significant interest in developing a digital twin for real-time cathodic protection diagnostics using Procor. This cutting-edge system allows Procor to function as a real-time monitoring tool capable of assessing corrosion risk, making it applicable across a wide range of systems.

Singing Capacitors Secrets Simulation

Presented By
Nopasorn Kowathanakul (Ansys)

Authored By
Nopasorn Kowathanakul (Ansys)
R. Sankar Kanamarlapudi (Ansys)

Capacitors are key components for the smooth operation of an electronic device by controlling the electrical flow throughout the PCB. Multilayer Ceramic Capacitors (MLCC) have a major role in modern electronic devices due to their small price and size, large range of capacitance, small ESL and ESR, and good frequency response. BaTiO3 dielectric is commonly used in MLCCs is known for its piezoelectric effect. When an electrical signal passes through the MLCC, the MLCC will contract and expand in proportion to the signal. This vibration of the MLCC will be transferred to the PCB to which it is soldered. The MLCC piezoelectric effect is often called œsinging . Simulating the singing secrets (piezoelectric characteristics, Vibration characteristics and Noise generation) requires a robust, accurate and comprehensive multi physics software like Ansys. This work demonstrates how œCoupled Field Harmonic  analysis system, latest practices of Acoustics preprocessing (advanced surface wrapper, PML extrusion layers) and comparison of FEM Acoustics vs BEM Acoustics in Ansys Mechanical environment. Piezoelectric material database available in Ansys Mechanical, makes it easy to select a standard piezoelectric material which has predefined matrices for anisotropic elasticity, piezoelectric and relative permittivity. With all these Multiphysics capabilities, developing a proper simulation methodology at the early design stage is preferable to address noise issues (through Harmonic Acoustics analysis) generated due to vibration characteristics (Modal analysis, Harmonic response). This work is more relevant to Semiconductors Industry R&D engineers to understand system level behaviour (Vibrations and Acoustics) of Semiconductor components.

Machine Learning Tools for Mutliphysics Analyses, Design Exploration and CAE Acceleration

Presented By
Nagananda Upadhya Gopalakrishna (BETA CAE Systems)

Authored By
Nagananda Upadhya Gopalakrishna (BETA CAE Systems)
Dimitrios Drougkas (BETA CAE Systems)

Machine learning tools have become increasingly prevalent in various software applications, offering numerous advantages and transforming the way we interact with technology. The challenge in the CAE-domain and Multiphysics analyses in particular, is to adapt and learn from the huge amount of Simulation data, plus demanding load cases and complex physics. The ML-workflow which involves data collection and preparation, ML-model training and evaluation, and finally Prediction, is integrated into the BETA Products since v21.0.0. The Prediction of scalar, 2d, 3d and Multiphysics Simulation results give us the opportunity to use Machine learning for variety of applications. The ML-based isolation of complex design components such as fasteners and clips can significantly reduce the modelling time. Predictors as the ML-object in BETA Products enable CAE engineers to rapidly explore various design configurations without relying on time-consuming simulations - a Multiphysics case will be presented of combining kinetics and durability simulations for automotive closures. Engineers can now optimize their processes, streamline workflows and expedite the realization of innovative designs. In this presentation we will present the latest features along with some of the most important applications.

AI-Powered Multiphysics Solutions: Revolutionizing the Design of Consumer Durable Systems

Presented By
Ujwal Patnaik (Altair)

Authored By
Ujwal Patnaik (Altair)
Nachiket Patil
Ismail Benhayoun (Altair)

The world of consumer electronics and durable systems is evolving at an extraordinary pace, propelling product designs into new realms of complexity. This complexity arises from the necessity to integrate a diverse array of physical phenomena seamlessly, aiming to meet the increasingly stringent demands for performance, reliability, and cost-efficiency simultaneously. Take, for instance, products as seemingly disparate as home audio speakers and washing machines: both encounter similar engineering challenges due to the convergence of various physical domains—mechanical, thermal, electromagnetic, and fluid dynamics (CFD). These challenges are further amplified by advancements in modern material manufacturing techniques. As consumer demands for smaller, more powerful, and multifaceted products increase, the need to solve multiphysics challenges during the design phase has become essential. The convergence of diverse physical phenomena, compounded by product miniaturization and complex system integration, defines a tightly constrained design arena. Here, conventional engineering methods frequently fail to meet the demands, making innovative approaches crucial for success. This paper explores how AI-powered engineering solutions offer a transformative approach to managing these design challenges. By leveraging AI, engineers can integrate complex physics models into one such method of reduced-order models that streamline the design process, allowing for real-time simulations, rapid design iterations, and optimization. AI enables the seamless coupling of diverse physics, facilitating the balancing of multiple design constraints (such as cost, durability, and performance) in ways that were previously unachievable. This approach reduces the need for extensive expertise in each domain while significantly accelerating product development timelines. We will present two case studies that demonstrate the utility and effectiveness of AI-powered multiphysics modeling: First Case Study: Home Audio Speaker Design: This product demands a multidisciplinary approach that balances structural integrity with acoustic performance, thermal management, electromagnetic compatibility, and aesthetic appeal. The speaker is designed to withstand structural vibrations, dissipate heat from its electrical components, and maintain consistent sound quality under various environmental conditions. Through the use of AI-powered optimization, engineers can simultaneously address these factors, ensuring the final design achieves performance objectives while remaining cost-effective and manufacturable at scale. Second Case Study: Washing Machine Design: Conversely, washing machines present a different set of engineering intricacies, with an emphasis on enhancing energy and water efficiency, fortifying material durability, and mastering moisture management. Optimizing the washing process involves fluid dynamics, thermal considerations, and robust control of mechanical components. Leveraging AI- powered multiphysics solutions, engineers can conduct precise simulations that delineate the dynamics of water and detergent distribution, energy utilization, and textile interactions. These simulations provide engineers with actionable insights, optimizing the appliance’s design towards superior efficiency and extended service life. Each case study demonstrates how AI-powered engineering can facilitate the convergence of multiple physics domains, allowing for a holistic approach to product design that accounts for the full range of physical interactions. The ability to solve complex multiphysics problems in a reduced time frame and with fewer manual interventions has profound implications for the future of product development in consumer durables. Through this paper, we aim to illustrate how AI-powered multiphysics modeling not only simplifies complex design problems but also pushes the boundaries of what is possible in engineering. By embedding AI into the design and simulation process, the development of next-generation consumer durable products becomes faster, more efficient, and more innovative. This shift ultimately results in higher-quality products that more adeptly meet the dynamic needs of consumers.

Accelerated and Reliable Predictions via PINN Incorporating Field Data

Presented By
Prashant Kumar Srivastava (oorja - weedy software)

Authored By
Prashant Kumar Srivastava (oorja - weedy software)
V. Dravid (oorja - weedy software)

In many scientific and engineering applications, accurate and efficient predictions are essential for decision-making processes. The integration of field data with physics-based models is an efficient strategy to enhance prediction accuracy while leveraging available observational information. This abstract proposes an approach employing Physics-Informed Neural Networks (PINNs) to achieve faster predictions by combining field data with underlying physical principles. The proposed methodology entails the following key steps: (1) Formulation of physics-based constraints derived from fundamental principles governing the system of interest, (2) Construction of a neural network architecture capable of approximating the system's response, (3) Training the PINN model using a combination of field data and physics constraints, and (4) Deployment of the trained model for quick predictions This abstract advocates for the adoption of PINNs as a versatile tool for accelerating predictions while harnessing the information contained within field data. Through illustrative examples for performance predictions on lithium ion batteries, we demonstrate the efficacy of this approach in enhancing predictive capabilities and enabling informed decision-making in complex systems.

Faster Design Space Exploration for Tires using Machine Learning Physics Model

Presented By
Klaus Krohne (Dassault Systemes)

Authored By
Klaus Krohne (Dassault Systemes)
Rasika Kulkarni
Girish Patil
Jayant Pawar
Atharv Govind
Jason Barr
Victor Oncea (Dassault Systemes)

Industry analysts forecast a nearly twofold increase in global vehicle miles traveled, driving a surge in tire demand which has led tire manufacturers to look for ways to remain competitive. The pivotal role of tire design in vehicle safety, comfort, and performance underscores the imperative for manufacturers to navigate complex trade-offs in key performance indicators (KPIs) while meeting stringent industry standards. Efficiently analyzing vast datasets to optimize tire designs is challenged by their inherent complexity. Despite the plethora of design optimization techniques available, the substantial time investment remains a significant hurdle. Leveraging advancements in data-driven technologies, particularly machine learning (ML), presents a promising avenue to expedite this design exploration process while conserving computational resources. For passenger car tires (PCR), maintaining a target footprint contact ratio across varying load conditions is desirable. Assessing the variation of footprint contact pressure contour and its sensitivity to tire maneuverability is another important KPI for the tire manufacturers. However, traditional nonlinear simulations entail long hours and resource-intensive computations for exploring the tire design space. To address these challenges, Dassault Systemes has developed a high-fidelity 3D parametric surrogate model to efficiently predict footprint contact ratio and contact pressure contour for diverse load and maneuverability scenarios, enhancing both speed and accuracy in tire design analysis. The methodology involves a comprehensive simulation setup, integrating a design of experiments (DOE) to evaluate input parameters'(namely the tread arc radii, bead filler height and vehicle load on the tire) influence on output parameters (namely the footprint contact ratio, contact pressure contour and the reaction forces on the road), alongside parallel training of a surrogate model using DOE-generated data. A comparative study demonstrates the surrogate model's fast and accurate predictions, facilitating assistance for KPI trade-offs. This approach is scalable across various tire segments and accommodating diverse KPIs like inflated diameter-section width, rolling resistance coefficient, wear resistance and the list goes on. The 3DEXPERIENCE platform offers a flexible framework to streamline tire design and virtual validation processes, empowering manufacturers to meet evolving market demands effectively.

FMU Validation with CosiMate: Using Native Co-Simulation to Prepare and Validate Multiphysics FMU-Based Simulation

Presented By
David Balland (Chiastek)

Authored By
David Balland (Chiastek)
C. Carrizo (ZF)
M. Jude (Alten)

FMI is a standard for co-simulation at system level. In this paper we will be presenting a strategy to co-simulate native and/or FMU models to create a proven FMU-based simulation. We will use CosiMate to support the activity of co-simulation. Our use-case is a Multi-Cylinder SI Engine, where GT model is representing the engine, Simulink model a controller and Amesim model a sample converter. The method for validation is iterative. We implement a first co-simulation with the original models thus creating a reference platform, then we export each model as an FMU for co-simulation, instantiate it in place of the native model and check the results against the reference. Co-simulation allows to use both native system models and/or FMU. this helps validating simulation workflow with less effort.. The co-simulation platform also allows multi-site distribution and we will discuss how this capability can offer the possibility to run a co-simulation using models located in different places or countries. In our paper we will discuss the benefits of native co-simulation to prepare FMU ™s; we will define some prerequisite to export and run an FMU using commercially available simulator; we will guide the integration of FMU into the co-simulation platform; finally, we will underline the advantages of FMUs for co-simulation and distributed co-simulation

Thermo-Chemical-Mechanical Modelling of Digital Light Processing 3D Printing

Presented By
Guang Liu (Institute of High Performance Computing (IHPC), A*STAR)

Authored By
Guang Liu (Institute of High Performance Computing (IHPC)
A*STAR)
David W. Rosen (Institute of High Performance Computing (IHPC)
A*STAR)

Digital light processing (DLP) is a type of vat photopolymerization additive manufacturing technology that has been widely used in practical applications due to its high printing resolution and fast printing speed. However, the distortions and residual stresses of DLP manufactured parts are significantly influenced by the complex thermo-chemical-mechanical coupled photopolymerization process during the layer-by-layer printing. Therefore, a comprehensive understanding of the system behavior is essential to predict the cure-dependent thermal and mechanical behavior of the layer-by-layer DLP process. In this study, a thermo-chemical-mechanical coupled modelling framework is developed to simulate the layer-by-layer DLP printing process by means of finite element analysis (FEA). Specifically, four sub-models are developed: irradiation model, photopolymerization model, thermal model and mechanical model. The curing dependent material properties are considered together with the shrinkage and heat generation caused by the chemical reaction. In addition, the progressive element activation method was implemented using the commercial FEA software ABAQUS to simulate the layer-by-layer photopolymerization in the DLP printing process. The release process to simulate the removal of the part from the build platform is also included in the framework. The capabilities of the proposed framework in predicting warpage and print-through effect have been verified with both 2D and 3D geometries from the literature.

Balancing Durability and MRI Compatibility: Evaluating Hip Implant Porous Designs

Presented By
Neha Parool (Dassault Systemes)

Authored By
Neha Parool (Dassault Systemes)
Himanshu Gururani
Abhishek Kadri
Sunil Sutar
Anand Pathak (Dassault Systemes)

Advancements in additive manufacturing has enabled realization of porous hip implants that possess a unique combination of durability, improved osseointegration and reduced osteoporosis. One of the major safety concerns related to these implants is risk of overheating in an MRI scanner. Multiple factors such as implant geometry, materials, radiofrequency and gradient coil design influence the implant heating. We present a novel multiphysics approach to explore the various designs of porous hip implant to arrive at an implant design with both in-vivo durability and MRI compatibility. For this purpose, multiple implant designs with different porous structures are created using lattice design application on the 3DEXPERIENCE platform and solved for in-vivo durability. Then, the porous implant geometries are imported in CST studio and anatomically positioned in the pre-existing in-silico human model corresponding to a 46-year-old male. Further, a 1.5T bird-cage MRI coil system in CST studio is used to evaluate maximum temperature rise in the implant and the specific absorption rate (SAR) in the surrounding tissues. The authors suggest that by leveraging the creative possibilities offered by additive manufacturing, complex porous structures can be strategically incorporated into implants, potentially balancing the durability and MRI safety. The present investigation encourages implant manufacturers to use multiphysics simulation models to optimize implant designs for improved durability and MRI safety.

Discussion Theme: AI Driven Methodologies in Multiphysics

Presented By
Alfred Svobodnik (MVOID)

Authored By
Alfred Svobodnik (MVOID)
X. Xiuqing (Institute of High Performance Computing (IHPC)
A*STAR)

Methods of Benchmarks for Multiphysics (Use Case Vibroacoustics)

Authored & Presented By
Alfred Svobodnik (MVOID)

Multiphysics & Quantum Computing

Authored & Presented By
Valtteri Lahtinen (Quanscient)