2024 RTC Technical Highlights Recap | Altair CAE and AI Technology Applications Conference

HyperWorks 2024 New Release Highlights

Richin Technology｜Chris

• HyperMesh 2024.1
  Continues to enhance FE-Geometry capabilities, extending more geometry editing functions (previously CAD-only) to FE-Geometry applications.
• SimLab 2024.1
  Further improves support for electromagnetic analysis setup, including automatic modeling templates for specific motor types, and significantly expands modeling capabilities for PCB and semiconductor 3D IC. It supports various analysis needs within HyperWorks in the same pre-/post-processing environment, lowering the learning barrier for users.
• OptiStruct
  Has implemented coupled explicit/implicit analysis capabilities. Altair is developing it into a main explicit/implicit finite element solver, so users only need to learn one solver to handle both explicit and implicit problems.
• PhysicsAI
  Expands support for more training parameters, including thickness, materials, and boundary conditions, enabling more complete AI modeling and prediction.
• ExpertAI
  Key features include machine-learning-based clustering and optimization analysis.
• Radioss
  Adds support for LS-DYNA keywords.
• HyperWorks
  The Python API is becoming more mature and now supports most custom development tasks, but it still cannot directly record macros into scripts, so users need learning resources to build scripts.

…Read the full article

Hypermesh2024 Latest Custom Development Techniques

(Supports Python API, co-development with Chat GPT)

Richin Technology｜Director Yu-Cheng Lin

Richin Technology Hypermesh2024 Custom Development Techniques

OptiStruct Comparison Applications with Third-Party Solvers

Quanta Computer Yi-Wen Chang, Manager

       OptiStruct is a full-featured structural solver. Using the HyperMesh convert function, Abaqus models can be converted into OptiStruct models almost seamlessly, with a conversion completeness of over 98%.
       Through real case comparisons (laptop open/close, fan operation simulation), Quanta found that OptiStruct results are very close to Abaqus results, with stress differences within 2%. The judgment of whether the product meets the analysis SPEC is also consistent with Abaqus, and for complex nonlinear problems it delivers comparable accuracy and reliability.

OptiStruct vs. Abaqus Stress Comparison

…Read the full article

Predicting Thermal Performance of an Automotive Liquid-Cooled Controller with physicsAI

Hon Hai Precision Industry｜Lead Engineer Chi-Shao Chen

      Hon Hai integrates Altair Inspire and SimLab, leveraging PhysicsAI technology to reshape the thermal design workflow and successfully overcome multiple challenges in automotive liquid-cooling thermal design.
       With parametric modeling and scripted batch modeling, design variants can be generated quickly, and PhysicsAI is used for performance prediction—truly enabling AI to replace CAE analysis. This approach requires only 10 seconds per prediction (180× faster than traditional methods), while maintaining high accuracy (MAE below 0.01 and confidence over 90%).
       In the future, PhysicsAI will be deeply integrated with design optimization technologies to further shorten development cycles and drive advances in thermal management.

 

Traditional FEA vs. physicsAI Prediction (Source: Hon Hai Precision Industry Co., Ltd.)

…Read the full article

AI-Driven Big Data Analysis for Air Conditioner Performance

Panasonic Taiwan｜Director Yi-Min Chen

Case 1:

       Using RapidMiner AI technology to automate processing of customer feedback. In the past, it took about three weeks to classify and summarize feedback. By training an AI classification model on historical data with RapidMiner, the deployment of the AI model significantly reduced the time required by relevant departments to process customer feedback.
 

Case 2:

       Analyzing historical IoT database data accumulated from air-conditioner usage, using clustering analysis to consolidate available datasets, then performing analysis to understand customer usage habits—with the goal of using AI to improve product and service quality.

Predicting Structural Strength of Heavy-Duty Hooks with physicsAI

G-Shock Enterprise｜Dr. Mien-Li Wang

Technical Highlights:

• Integration with third-party solvers
• Supports multiple training parameters: boundary conditions, materials, geometry, etc.

       G-Shock Enterprise adopts Altair PhysicsAI technology, integrating historical design and analysis data to train and build an AI prediction model for rapid prediction of hook structural strength.
Through HyperMesh  for parametric design and DOE analysis in HyperStudy , an efficient AI prediction workflow was established.
       PhysicsAI The model can predict hook structural strength within seconds, with only about 3% error compared to traditional CAE analysis. Compared with hours of CAE computation, the AI model significantly shortens the design cycle, improves efficiency and accuracy, demonstrates AI’s application potential in engineering design, and successfully drives an intelligent transformation of the design workflow.
 

Comparison of CAE vs. PhysicsAI Prediction for Heavy-Duty Hooks

Gap Analysis and Optimization of Rocket Payload Fairing Locking Method

National Space Organization｜Associate Engineer Ting-Wei Chen

       Rocket structure lightweighting and cost control are key design priorities. Among them, the payload fairing, which protects the payload, is especially critical.
       This project combines Altair Inspire  and HyperStudy , uses computational fluid dynamics (CFD) for aerodynamic pressure simulation, and supplements it with internal/external pressure differential analysis, thermal protection design, and rigid-body motion simulation to build a complete analysis workflow. By setting seven design variables (DV1–DV7) and applying the Global Response Surface Method (GRSM) for multi-objective optimization, the design reduced the fairing gap to 0.21 mm and successfully reduced the structural weight by 4 kg.
       This optimization improves payload capability while also saving approximately USD 150,000 in launch cost, achieving an optimal balance between performance and cost.
 

Fairing Displacement Contour (Source: National Space Organization)

Design & Analysis of Magnetic Components for High-Voltage, High-Power Converters

ITRI Green Energy & Environment Research Laboratories｜Researcher Frank Nguyen

       Energy storage systems require high-power, high-voltage inductive components, and thermal loss and temperature rise considerations are key factors during design. Using Flux 3D  to assist in inductor design and analysis not only enables simulation of magnetic fringing flux and calculation of losses such as core loss and copper loss, but also further enables temperature-rise prediction for magnetic components at thermal steady state.
       This project uses Flux  to evaluate the overall impact of different design parameters on inductor performance, reducing time and cost spent on building physical prototypes.
 

Flux3D Inductor Magnetic Flux Density Vector Plot

PCB and Chassis Structure Multi-Stage Assembly Simulation (Maximizing HyperWorks Capabilities)

Richin Technology｜Frank Su

Solder Fatigue Analysis of Package Solder Balls

Richin Technology｜William

Server Fan Aerodynamic Noise Simulation – GPU-Accelerated ultraFluidX

Richin Technology｜Jason

Easily Complete Magnetic Analysis for Consumer Electronics

Richin Technology｜Peggy

Perfect Integration of SimLab and Flux

Richin Technology is an “expert in CAE and AI data analytics,” and we have delivered many successful case studies.

▶ Contact us now to get more information.
▶ Subscribe to the Richin YouTube channel to explore more CAE and data analytics.