The latest release of CAESES introduces major enhancements across usability, system integration, security, and geometry handling. Version 2026.1 continues to strengthen performance, interoperability, and overall design workflow efficiency.

New release naming

With version 2026.1, CAESES introduces a new naming scheme and rolling release strategy. This approach supports a more continuous delivery model, enabling regular improvements while maintaining clear version traceability across updates.

Dark Mode & Refreshed Interface

CAESES now features a modernized graphical user interface and a new dark mode option. The updated design improves consistency, readability, and overall usability across the application.

Key improvements include:

• Improved visual consistency across modules
• Enhanced readability in complex modeling environments
• More comfortable long-session usage
• Refined layout behavior for flexible workspace handling

View full image

Windows Desktop Integration

CAESES is now more deeply integrated into the Windows environment, improving everyday usability and workflow continuity.

Key updates include:

• Native support for Windows Snap Layouts
• Standard Windows keyboard shortcuts
• Smoother window resizing and docking behavior
• Improved multitasking across applications

View full image

Project-Level Security & Access Controls

A wide range of security capabilities in the new add-on Advanced Project Security provide stronger control over sensitive design data at project level.

Key features include:

• Password or license-based project locking
• Restriction of feature editing rights
• Clipboard extraction protection for sensitive models

View full image

Plug In Your Own Optimization Algorithm

CAESES now enables seamless integration with external optimization workflows via the new External Optimization Engine.

This allows users to:

• Connect external optimization algorithms directly to CAESES
• Drive geometry generation from third-party tools
• Exchange data via a structured JSON interface

Native Parasolid export

Parasolid export is now fully integrated into CAESES Standard Edition. Completely re-engineered as a native module, it replaces the previous translator with an in-house implementation that is maintained and developed independently of external release cycles.

Key updates include:

• No separate add-on license required
• Support for ASCII (.x_t) and binary (.x_b) formats
• Improved robustness for complex geometries
• Fully in-house implementation, independent of external translators

A fresh look for CAESES

CAESES now comes with a renewed visual identity; think of it as a fresh face for your everyday design environment.

The update includes a new logo, a refined color scheme, and a more consistent visual language across the platform and communication materials. The aim is a cleaner, more modern look that improves recognition and brings a more unified experience across all touchpoints.

What else?

This release also includes a range of additional improvements, refinements, and fixes, to know more about all of them, see the Changelog.

The 2026 CAESES North America User Conference marked an important milestone as the first conference held specifically to better serve and support the North American user community. While the event was focused on the NA community, it was wonderful to see several users join from outside the region as well. As a virtual event, the conference brought together a diverse group of attendees, presenters, and sponsors, and we are deeply grateful to all of them for making this inaugural event a success.

Presentations

The conference featured an impressive range of presentations spanning multiple industries, reflecting the breadth and diversity of the CAESES user community. The Q&A sessions following each presentation were particularly insightful and added great value to the discussions.

Below you will find links to the slides and recordings for each session:

Looking back and looking forward: Simulation-driven Design and Data-driven Engineering
Stefan Harries, Managing Director, FRIENDSHIP SYSTEMS AG
Slides | Recording 

Outlet Volute Optimization of the Dragon Heart, a Total Artificial Heart for Pediatric Patients
Giselle Matlis, PhD candidate, Drexel University, School of Biomedical Engineering
Slides | Recording 

Structured Meshes Powering Simulation-Driven Design
Samuel James, Chief Operating Officer (Global Operations), GridPro
Slides | Recording 

The Importance of Hydrodynamic Optimization in the Development of Small Electric Marine Vessels
Britton Ward, President, Farr Yacht Design
Slides | Recording 

Engine Combustion System Design with CAESES: A Passive Pre-Chamber Case Study
Anqi Zhang, Senior Research Engineer, Aramco Americas: Aramco Research Center – Detroit
Slides | Recording

Optimization of Hydrokinetic Turbine Blade
Dae-Hyun Kim, Senior Engineer, Technology Americas, American Bureau of Shipping (ABS)
Slides | Recording

Performance Evaluation of Pump Volute Geometries Optimized for Additive Manufacturing
Thiago Ebel, Director of Technology and Innovation, ConceptsNREC
Slides | Recording

Hydrodynamic Optimization of Surface Combatant using Multi-Fidelity Co-Kriging algorithm
Giacomo Pellizzari, PhD Candidate in Ocean Engineering, Kevin T. Crofton Department of Aerospace and Ocean Engineering, Virginia Tech
Slides | Recording

Design by Fully-Parametric Optimization of Small AUVs Ducted Propellers
Samiksha Dhakal, PhD Candidate in Aerospace Engineering, Kevin T. Crofton Department of Aerospace and Ocean Engineering, Virginia Tech
Slides | Recording

Workshops

A special thank you to Simon and Andreas for organizing and leading the hands-on workshops. Both sessions were well-received and gave attendees the opportunity to deepen their practical knowledge of CAESES.

CAESES workshop 1: morphing

This workshop focused on morphing workflows in CAESES and demonstrated how geometry variations can be efficiently created and controlled within simulation-driven design processes. Participants gained practical insights into flexible geometry manipulation techniques and their application in engineering workflows.

Watch Workshop Recording

CAESES workshop 2: automation and integration

This workshop explored automation and integration workflows within CAESES, highlighting how engineering processes can be streamlined through connected toolchains and automated design studies. The session showcased practical approaches for improving workflow efficiency and scalability across simulation environments.

Watch Workshop Recording

Future Events

Looking further ahead, we plan to hold the next CAESES North America User Conference in 2028

We hope to see more of you soon!

Developing a geometry that fulfills all given requirements – such as due to performance and manufacturing – is rarely a straightforward process. In many engineering projects, especially those involving aerodynamic surfaces, lightweight structures, or customized components, the challenge extends far beyond building a 3D model.

The geometry model must be robust when changing parameter values, consider manufacturing constraints, integrate with simulation workflows, and generate suitable export data for following processes and production. Even relatively small adjustments can affect multiple downstream components, making flexibility and workflow stability increasingly important throughout development.

To better understand how different CAD tools approach handle these challenges, a comparison study was carried out on the vertical stabilizer of the human powered aircraft “Libelle” from Odonata e.V. as a practical engineering example.

View full image

The goal was to create high-performing and manufacturing-ready geometries, and more specifically, export DXF files suitable for laser cutting and hot wire cutting processes while simultaneously maintaining enough geometric flexibility to support future design changes.

The project also provided an opportunity to compare how CAESES and conventional CAD software perform in a real component design workflow.

The components

The stabilizer consists of several interconnected components manufactured using different materials and processes.

While generating these components, several parameters needed to remain adjustable throughout development, including:

• Airfoil geometry
• Chord length at root and tip
• Thickness distribution
• Sweep
• Geometric angle of attack

This created a highly iterative workflow where geometry changes can influence multiple downstream components simultaneously.

Two different CAD approaches

Conventional CAD software and CAESES approach geometry generation differently.

Traditional CAD systems are typically built around history-based modeling workflows. Features are created sequentially, with each operation depending on previous geometry definitions. This approach works well for finalized production models, technical drawings, and assembly-focused workflows.

However, as geometry becomes more complex and iterative, and updates become more frequent, history-based workflows can become increasingly difficult to manage. Larger modifications often create broken references, unstable features, or additional manual remodeling work.

CAESES approaches the problem from a more parametric perspective. Instead of focusing primarily on sequential feature histories, CAESES builds geometry around parameters, relationships, and automated dependencies. This allows engineers to modify critical design parameters while maintaining stable downstream geometries and manufacturing outputs.

For manufacturing-oriented projects with evolving geometries, this creates a fundamentally different workflow experience.

View full image

Building the geometry in CAESES

The geometry in CAESES was created using a combination of points, curves, surfaces, Breps, and Boolean operations. The overall workflow was intentionally kept compact by minimizing the number of controlling parameters while still maintaining enough flexibility for future design modifications.

Airfoil sections form the foundation of the stabilizer geometry. Parameters such as chord length, thickness, spar position, and spanwise positioning can directly be adjusted within the model definition. The spar position plays an important role because it influences the aerodynamic control behavior of the aircraft.

Using ruled surfaces between the airfoil sections, the stabilizer geometry could efficiently be generated while remaining fully parametric throughout development. One important advantage of this approach is that geometry modifications can be applied at virtually any stage without requiring large portions of the model to be rebuilt manually.

The resulting surfaces were converted into watertight Brep geometries suitable for downstream processing such as Boolean operations, simulation preparation, and manufacturing export.

As the workflow expanded, folders In CAESES were used to organize the individual manufacturing components, including ribs, molds, templates, and structural supports. This made it easier to manage increasingly complex relationships between components while maintaining a clear workflow structure.

The manufacturing geometries themselves were generated through Sub-Breps and Boolean operations. Rib sections, for example, can be created automatically based on spacing definitions and rib thickness parameters.

Instead of manually positioning each component individually, parametric relationships control the placement and generation process. This significantly reduced repetitive modeling work and simplified later geometry modifications.

Finally, the individual components were exported as DXF files suitable for laser cutting and hot wire cutting systems.

Comparing the workflows

The same geometry was also created using a conventional CAD tool by an experienced engineer familiar with both systems. The comparison focused primarily on overall engineering effort and workflow robustness rather than simply comparing feature lists. One important observation was the difference in how both systems handled geometry changes.

Within the conventional CAD workflow, the history-based modeling structure introduced additional manual work whenever larger geometry modifications affected downstream features. As dependencies became more complex, rebuilding and repairing geometry relationships required increasing amounts of time.

In CAESES, the parametric workflow structure handled these updates much more efficiently. Because relationships between components are embedded directly into the model logic, many geometry changes propagated automatically throughout the workflow. This reduced manual remodeling effort significantly and helped maintain stable manufacturing outputs even as the geometry evolved.

Development time comparison

The difference between the two approaches became particularly visible in the total development time.

Creating the complete manufacturing-ready geometry required:

• approximately 6 hours in CAESES
• approximately 14 hours in a conventional CAD tool

The difference became even more relevant once geometry modifications were introduced. While the conventional CAD workflow required additional troubleshooting and rebuilding effort, CAESES maintained a significantly more stable workflow structure during iterative updates. This was one of the clearest indicators of how parametric geometry generation can improve efficiency in manufacturing-ready engineering projects.

Where CAESES shows clear advantages

Robust geometry updates

One of the strongest advantages observed in CAESES is the ability to apply geometry changes without destabilizing the workflow.

In the given example, parameters such as:

• Airfoil definitions
• Chord lengths
• Thickness distributions
• Sweep
• Structural positioning

could be adjusted while maintaining functional downstream geometries and manufacturing outputs.

For iterative engineering projects, this level of robustness can significantly reduce manual rework.

Efficient airfoil geometry creation

Creating aerodynamic geometries using standard definitions such as NACA series or CST curves is straightforward within CAESES.

This simplifies the setup of airfoil-driven geometries and reduces manual preparation work.

Simulation-ready geometry

The generated geometry is also immediately suitable for simulation workflows, including:

• Structural analysis
• Aerodynamic analysis
• Low-fidelity simulation methods
• High-fidelity CFD workflows

This reduces the need for additional geometry preparation before analysis.

Geometry checking and visualization

CAESES also provides direct visualization of problematic geometry areas such as open edges.

This improves geometry reliability during modeling, export preparation, and simulation setup.

Strong support for free-form geometry

The flexibility of the CAESES modeling approach is particularly effective for highly customized and free-form geometries.

Custom feature definitions and parametric relationships make it easier to maintain adaptable workflows without becoming constrained by rigid feature histories.

View full image

Conventional CAD in traditional workflows

Conventional CAD systems remain widely used for production drawings, assemblies, and standard documentation workflows. Their interfaces and workflows are also highly familiar across the engineering industry.

However, this comparison highlights the limitations of history-based modeling when handling iterative geometry changes, automated manufacturing preparation, and highly parametric workflows.

As geometry complexity increases and updates become more frequent, CAESES maintains a significantly more stable and efficient workflow structure throughout the project.

Final thoughts

The comparison demonstrated that both CAESES and conventional CAD software are capable of generating manufacturing-ready geometry.

However, the workflows differ significantly once geometry flexibility, iterative updates, simulation integration, and automated manufacturing preparation become important requirements.

Conventional CAD systems remain highly effective for production-focused engineering tasks and standardized documentation workflows.

CAESES demonstrated clear advantages in:

• Parametric geometry generation
• Robust handling of design changes
• Automated manufacturing preparation
• Simulation-ready geometry workflows
• Free-form geometry modeling

For engineering teams working on aerodynamic surfaces, lightweight structures, or highly iterative component manufacturing projects, these advantages can translate directly into reduced manual effort, faster geometry updates, and more efficient development workflows.

As engineering processes continue to become more simulation-driven and data-centric, robust parametric geometry workflows are becoming increasingly important across advanced design applications.
 

Designing an efficient aviation propeller is far more complex than it first appears. What seems like a simple rotating component quickly turns into a highly sensitive engineering challenge, where aerodynamics, structural constraints, and performance requirements are tightly interconnected. Even small geometric adjustments can significantly influence thrust, efficiency, noise, or vibration behavior.

Despite this complexity, many design workflows still rely on repetitive, manual steps – rebuilding geometry, rerunning simulations, and retracing paths that have already been explored. This not only slows down development but also limits how thoroughly designers can investigate alternative concepts. A more integrated, parametric approach offers a clear advantage by connecting geometry, simulation, and optimization into a continuous workflow.

View full image

Why aviation propellers are so challenging

Although aviation and marine propellers share similar principles, operating in air introduces very different physical challenges. At higher rotational speeds, compressibility effects become relevant, and parts of the flow can enter transonic regimes, potentially causing shock waves and efficiency losses. In addition, performance is highly sensitive to Reynolds number variations along the blade.

Unlike marine applications, where cavitation is a primary concern, aviation propellers must meet strict requirements for noise and vibration. These constraints, combined with changing flow conditions along the blade radius, result in a system where every geometric detail matters, and simplifications can quickly lead to suboptimal designs.

Moving beyond manual geometry

Traditional propeller design workflows often involve direct manipulation of geometry – adjusting surfaces, exporting models, repairing issues, and repeating the process. While this approach can work for simple iterations, it becomes increasingly inefficient as complexity grows.

A parametric modeling strategy fundamentally changes how geometry is handled. Instead of editing shapes directly, the designer defines relationships that govern how the propeller is constructed. Parameters such as chord length, pitch angle, thickness distribution, and blade curvature are described mathematically and linked together.

This means that when a single parameter is adjusted, the entire geometry updates automatically and consistently. There is no need to rebuild the model or fix broken surfaces. The propeller remains simulation-ready at all times, allowing for rapid iteration and exploration.

More importantly, this approach encourages a deeper understanding of the design itself. Rather than focusing on isolated geometry tweaks, engineers can think in terms of cause and effect – how changing a distribution or parameter influences overall performance.

View full image

One workflow, different propeller concepts

Aviation propellers come in many forms, from simple fixed-pitch designs to more advanced variable-pitch systems. While the underlying physics remains consistent, the design priorities and trade-offs can differ significantly depending on the application.

Fixed-pitch propellers are valued for their simplicity, low weight, and reliability. They are commonly used in unmanned aerial vehicles, ultralight aircraft, and training platforms where robustness and ease of use are essential. However, their performance is inherently limited to a narrow operating range.

Variable-pitch propellers, on the other hand, introduce additional flexibility by allowing the blade angle to change during flight. This enables better performance across different phases such as takeoff, climb, and cruise. The added complexity, however, requires more careful design and integration.

A parametric workflow makes it possible to explore both concepts within the same overall framework. Designers can even switch between configurations, compare performance, and evaluate trade-offs without starting from scratch each time. This significantly reduces development time and opens the door to more comprehensive design studies.

From fast estimates to high-fidelity analysis

Simulation plays a central role in propeller design, but not all methods are equally suited for every stage of development. High-fidelity CFD provides detailed insights, but it is too time-consuming for early-stage exploration.

This is where faster methods, such as blade element models, become valuable. These methods provide quick estimates of key performance metrics like thrust, torque, and power consumption. While less detailed than CFD, they are invaluable for screening concepts and identifying promising directions.

By combining fast, approximate methods with selective use of high-fidelity simulations, engineers can strike a balance between speed and accuracy. This layered approach ensures that computational resources are used where they provide the most value.

View full image

Integrating simulation into the process

No matter which tool is used at the respective stage of the process, an integrated workflow truly proves its worth. Instead of manually preparing simulation models, engineers can automate the generation of computational domains, mesh refinement regions, and solver setups.

This not only reduces preparation time but also improves consistency and reproducibility. It even allows designers to expand the scope of their analysis beyond the propeller itself.

In real-world applications, a propeller does not operate in isolation. The surrounding aircraft geometry – fuselage, cockpit, landing gear, and structural components – affects the airflow and, ultimately, the propeller’s performance. Ignoring these interactions can lead to misleading conclusions.

An integrated approach makes it feasible to include these effects in the simulation process, leading to more realistic and reliable results.

A real-Life example: the “Libelle” human-powered aircraft

The importance of system-level thinking becomes particularly clear in innovative projects such as the “Libelle” human-powered aircraft developed by the student team Odonata e.V. in an effort to break the world record for the longest distance flight under human power. In such an extreme design scenario, efficiency margins are incredibly tight, and even small aerodynamic interactions can have a noticeable impact.

View full image

Rather than optimizing the propeller in isolation, the development team considered the entire aircraft as a coupled system. By integrating the cockpit and supporting structures into their analysis, they were able to capture interference effects that would otherwise have been overlooked. 

View full image

This holistic approach enabled them to better understand how airflow behaves around the aircraft, reduce performance losses caused by interactions, and compare multiple design options under realistic conditions. The result was not just a better propeller but a more optimized overall system. 

View full image

View full image

View full image

What are we actually optimizing for?

At its core, propeller design is an exercise in balancing competing objectives. Improving one aspect of performance often comes at the expense of another, and there is rarely a single “perfect” solution.

Designers must consider efficiency across a range of operating conditions, not just a single design point. They need to ensure that sufficient thrust is generated while keeping torque and structural loads within acceptable limits. At the same time, noise and vibration must be minimized, and weight should be kept as low as possible.

These competing requirements make it essential to adopt an iterative and flexible design process. Rather than searching for a one-step solution, engineers must understand the design space to continuously refine and evaluate their designs, guided by both data and experience.

Where this approach fits

The benefits of a connected, parametric workflow are not limited to a specific type of aircraft. They apply broadly across the aviation industry and beyond.

From general aviation and unmanned aerial systems to electric propulsion concepts and experimental research projects, the ability to quickly explore design variations and base decisions on performance data is increasingly valuable.

As propulsion technologies evolve and new requirements emerge, the need for adaptable and efficient design processes will only continue to grow.

Toward a more efficient design process

Ultimately, the shift in propeller design is centered around integration. When geometry generation, performance analysis, and optimization are all part of a unified workflow, the entire development process becomes more efficient.

Engineers spend less time rebuilding models and more time improving them. They gain clearer insights into how design choices affect performance and can respond more quickly to new challenges or requirements.

Starting a new engineering project should be exciting. But for many engineers, the first days, or even weeks, are spent rebuilding models, recreating workflows, and rediscovering solutions that already exist.

Whether you’re part of a large team, a small firm, or even working solo, this problem is the same: valuable time is wasted, mistakes are repeated, and the opportunity to explore better solutions is lost.

This isn’t a matter of skill, but rather how the engineering work is structured.

Even individual engineers feel the drag. Working from scratch on every project can be frustrating, inefficient, and demotivating. And when multiple engineers or teams are involved, the problem compounds, slowing down progress and innovation across the board.

View full image

Why starting from scratch holds engineers back

Rebuilding every project from zero comes with costs that affect both individual engineers and teams:

1. Lost knowledge

Past solutions, tweaks, and lessons often exist only in an engineer’s mind. When a project ends or when someone moves on, that knowledge disappears. Starting over means rediscovering solutions to problems that have already been solved.

2. Repeated errors

Without standardized practices, small mistakes keep appearing project after project. For individual engineers, this can be frustrating; for teams, it leads to inconsistent results.

3. Wasted time

Time spent on reconstruction is time not spent innovating or improving designs. Individual engineers may feel stuck in routine work rather than applying their skills to something more meaningful.

4. Inconsistent outcomes

Every new project can feel like reinventing the wheel. The results may vary widely depending on who’s handling it, making it harder for engineers to feel confident in their decisions.

View full image

How engineers can work smarter

Whether you’re a part of a large team or working independently, there are practical ways to avoid starting from scratch:

1. Parametric models

Create models that can adapt to new requirements. A single model can serve as a template for multiple projects, reducing repetitive work while letting engineers focus on innovation.

2. Reusable workflows

Standardize recurring tasks, simulation setup, validation, or testing, so you don’t have to redo the same work each time. Even small routines can make a big difference in productivity.

3. Capturing knowledge

Document processes, solutions, and lessons learned. Individual engineers benefit by having a personal “playbook” to speed up future projects, while teams benefit from shared knowledge.

4. Exploring multiple options

With structured models and workflows, engineers can test multiple design alternatives efficiently, uncovering better solutions without extra effort.

Even small improvements to the workflow can transform an engineer’s daily work from repetitive to meaningful, giving time to focus on creativity and problem-solving.

View full image

Where CAESES helps both engineers and teams

CAESES is designed to make parametric, reusable, and automated workflows practical for both individuals and teams. It allows engineers to:

• Reduce repetitive tasks without sacrificing flexibility
• Reuse models and processes across multiple projects
• Explore design alternatives systematically
• Capture knowledge for future use

For individuals, this means less frustration, fewer repetitive tasks, and more time for creativity. For teams, it ensures everyone benefits from shared knowledge and standardized practices.

These workflows also provide a foundation for data-informed decision-making, allowing both engineers and teams to make smarter choices faster.

View full image

Why early setup matters

A common lesson from high-performing engineers is that the first steps of a project matter more than most people think. How you set up your model, workflow, and data at the start determines how efficiently you can iterate and explore options later.

Even solo engineers benefit from well-structured workflows that let them reuse their own past work and scale their efforts. In teams, consistent early setup ensures everyone works from the same foundation, reducing errors and miscommunication.

Beyond efficiency: confidence and innovation

Starting from scratch is more than an inconvenience, it’s a bottleneck that slows engineers of all levels and limits the potential of entire teams. By creating reusable models, structured workflows, and systems for knowledge capture, engineers can work faster, make smarter decisions, and focus on innovation.

CAESES make this possible, providing tools that empower engineers to work efficiently, reduce repetition, and build a foundation for better decision-making.

The future of engineering isn’t about starting over. It’s about building on what you’ve already achieved, whether you’re working alone or with a team, and doing it smarter.

View full image

Fitting Existing Hull Forms

The Component-Based Ship Workflow in CAESES provides a structured method for creating parametric ship hull geometries for modern ship design and naval architecture. This approach also enables the parametric fitting of benchmark ship hulls developed by the Korea Research Institute of Ships and Ocean Engineering (KRISO), including the KCS (KRISO Container Ship) and KVLCC2 (KRISO Very Large Crude Carrier). These hulls provide high-quality datasets widely used for CFD validation, hydrodynamic analysis, and ship flow studies. Parametric versions of these models are available in the CAESES sample library for simulation-driven ship design and optimization.

Comparison Between Original and Parametric Geometry

A comparison of the station curves between the original hull geometry (red) and the parametric reconstruction created with the Ship Modelling Workflow (green) demonstrates the accuracy of the fitting process while enabling parametric control for further hull form exploration and optimization.

View full image

View full image

A comparison of the hydrostatics between the original geometry and the parametric ship modeling workflow (SMW) model is also performed.

Table 1: Hydrostatics Comparison for KCS

Table 2: Hydrostatics Comparison for KVLCC2

From KCS to KVLCC2

Both parametric hull models use the same Ship Modeling Workflow (SMW) parameterization in CAESES, including:

• Section-Based Aftship
• Waterline-Based Foreship
• Single Skeg
• Basic Bulb

Because both hulls are generated using the same parametric modeling framework, they belong to the same parametric hull design family, even though they represent different ship types. The KCS is a container ship, while the KVLCC2 is a tanker, resulting in significant differences in hull fullness, proportions and cargo capacity requirements.

Due to the shared parameterization structure, however, the SMW model enables a smooth geometric transition between the two hull forms. By adjusting a single design variable, the geometry can continuously morph from the KCS hull form to the KVLCC2 hull form. This demonstrates the flexibility of parametric ship hull modeling and shows how a unified parameterization can represent a wide range of ship hull designs within one parametric framework.

View full image

View full image

In the animations, ship appendages including the rudder and propeller are also incorporated to provide a more realistic representation of the propulsion configuration. The propeller geometries are based on designs created using the Advanced Propeller Workflow in CAESES and imported into the project as STEP files for simulation. The KCS uses the 5-bladed KP505 propeller, while the KVLCC2 is equipped with the 4-bladed KP458 propeller, both designed within the same parametric workflow before integration into the model. The radial distributions of the propeller design parameters for both propellers are shown below, highlighting the key geometric characteristics of the two propeller configurations.

View full image

Deformation

Based on these reference models, the identified parameters of the Ship Modeling Workflow parametric hull model were defined as the baseline geometry. This baseline served as the starting point for additional targeted shape modifications, introducing further geometric flexibility to analyze the hydrodynamic impact of local hull variations. Two shape deformation techniques were applied.

First, free-form deformation was used on the KCS bulbous bow, enabling smooth and controlled bulb shape variations while preserving surface continuity. This allows systematic exploration of alternative bulbous bow designs without compromising hull geometry quality. Second, RBF-based B-Rep morphing was applied to the aftbody of the KVLCC2 hull, enabling precise and localized stern geometry modifications directly on the CAD model while maintaining high geometric fidelity.

Together, these methods demonstrate how partially parametric deformation techniques can be applied to ship hull models to enable targeted design exploration and hydrodynamic optimization within a simulation-driven design framework. The animations below illustrate the influence of the four design variables used in each partially-parametric model.

View full image

View full image

Simulations with SHIPFLOW

SHIPFLOW is a Computational Fluid Fynamics (CFD) software specialized in the hydrodynamic analysis of ships and marine propellers. It is widely used in naval architecture for predicting resistance, propulsion performance, and flow characteristics around ship hulls. CAESES is well connected with SHIPFLOW, enabling seamless integration between parametric geometry modeling and hydrodynamic simulation. Through this connection, hull forms created and modified in CAESES can be directly transferred to SHIPFLOW, allowing efficient design exploration, automated simulations, and optimization of ship performance.

KCS Setup – Resistance Simulation

The original KCS design point corresponds to a design draft of 10.8 m and a service speed of 24 knots, conditions for which the bulbous bow geometry was optimized. This configuration serves as the baseline hull design for evaluating hydrodynamic performance. To assess performance across different operating conditions, additional scenarios were considered: a reduced draft of 9.5 m representing lighter loading, slow steaming at 18 knots for energy-efficient operation, and a high-speed condition of 26 knots.

Hydrodynamic simulations were performed using XPAN from SHIPFLOW, a potential-flow solver widely used for ship resistance prediction and hull-form evaluation. A fine computational mesh was applied to capture the flow characteristics around the hull and bulbous bow region. The analysis focused on key resistance components, including frictional resistance (Rf) and wave-making resistance (Rw), providing insight into the vessel’s hydrodynamic performance across operating conditions.

KVLCC2 Setup – Self-Propulsion Simulation

The KVLCC2 baseline configuration corresponds to a design draft of 20.8 m and a service speed of 15.5 knots. The study focused on self-propulsion simulations to evaluate the interaction between the ship hull and propeller. Simulations were conducted using the XCHAP from SHIPFLOW, a RANS solver that combines propeller–hull interaction modeling with propulsion analysis. Calculations were performed at model scale using a coarse mesh, enabling the evaluation of multiple design variants with reasonable computational cost.

Propulsion was modeled using a Wageningen B-series propeller, a widely used reference propeller series in naval architecture and hydrodynamic studies. The evaluation focused on key propulsion metrics, including total resistance (RS), power delivered (PD), and total propulsive efficiency (ηDS). These parameters provide a comprehensive assessment of propulsion performance and energy efficiency.

Design of Experiments (DoE)

For each partially-parametric model, four design variables controlled the geometric modifications. The Sobol sequence sampling method was used for design space exploration, providing well distributed coverage of the parameter space. Following common DoE guidelines, typically 5 to 10 samples per design variable, a total of 30 design variants were generated for each study. This sampling strategy provides a sufficiently rich and well-distributed dataset for the development of surrogate models, enabling the training of reliable machine learning models to approximate the underlying simulation responses.

KCS Resistance (DoE)

Resistance simulations were performed at 18 knots and 26 knots to evaluate hull performance under different operating conditions. For each of the 30 sampled designs, flow-field visualizations were analyzed to understand the hydrodynamic effects of geometric variations. The animations illustrate pressure distribution, streamlines, and wave elevation, providing insight into how bulb geometry modifications influence flow behavior and ship resistance across the two operating scenarios.

View full image

View full image

During post-processing, correlation and regression analyses in CAESES were used to evaluate how the design variables influence hydrodynamic performance. These analyses help identify parameter sensitivities and trends, revealing how each variable affects ship resistance within the explored design space.

A key observation emerges when comparing results at the two operating speeds. Changing speed significantly alters the influence of the design variables on wave-making resistance, showing that parameter importance varies across operating conditions. This effect is also visible in the animations. At 18 knots, the results favor a shorter and more slender bulbous bow, which helps reduce wave resistance during slow steaming. At 26 knots, the design exploration leads to a longer and fuller bulb, better suited for higher-speed operation and different wave patterns.

View full image

View full image

View full image

KVLCC2 Self-Propulsion (DoE)

Sobol sampling was used to generate 30 design variants based on four design variables, ensuring a well-distributed design space exploration. The animations illustrate pressure distribution, streamlines, and nominal wake for the different designs. These visualizations help analyze how stern geometry variations influence hull flow, propeller inflow and overall propulsion performance, providing insight into the vessel’s hydrodynamic and propulsive efficiency.

View full image

View full image

Correlation and regression analyses were carried out during the post-processing stage using CAESES, allowing the influence of the design variables on the hydrodynamic performance to be systematically evaluated. The analysis indicates that a wider skeg in the lower region tends to increase the overall resistance of the vessel. However, at the same time, an increase in total propulsive efficiency (ηD) is observed. In several cases, the improvement in propulsive efficiency is larger than the corresponding increase in resistance, resulting in an overall reduction in the required propulsion power.

This behavior is partly related to the characteristics of the ITTC extrapolation method, which is commonly used for resistance and propulsion analysis at model scale. While the method provides consistent comparative trends for design evaluation, the exact balance between resistance and propulsion efficiency improvements may not fully represent the behavior at full-scale operating conditions.

View full image

Conclusion

This study demonstrates the capabilities of fully parametric modeling using the Ship Modeling Workflow in CAESES, successfully reproducing existing hull forms such as KCS and KVLCC2. Partially-parametric modeling was also applied, highlighting the robustness and efficiency of parametric approaches.

A seamless integration between CAESES and SHIPFLOW enabled an automated simulation-driven design framework. Resistance simulations were performed using the XPAN solver, while self-propulsion analyses were conducted with XCHAP.

The results highlight the power of simulation-driven design, where numerical simulations combined with parametric modeling and structured design exploration reveal complex relationships between hull geometry and hydrodynamic performance. This approach enables designers to identify optimal design trends and make informed decisions early in the design process.

The study also emphasizes the increasing importance of data-driven engineering. Managing and analyzing large volumes of simulation data allows meaningful insights to be extracted, supporting smarter design processes and more efficient development of high-performance ships. The integration of Artificial Intelligence (AI) and machine learning techniques further enhances this process by enabling the development of predictive models that can approximate simulation results, identify complex patterns within the design space, and support faster design exploration and optimization.

Overall, the combination of parametric modeling, simulation-driven design, and data-driven engineering provides a powerful and structured workflow for modern ship design, enabling efficient progression from concept development to optimized hull performance.

Engineering teams today face increasing pressure to deliver high-performance designs faster and with fewer resources. While advanced simulation tools like CFD and FEM are widely used, many teams still struggle with fundamental challenges in parametric modeling, optimization, and design automation.

In this blog, we highlight five common engineering challenges that slow down simulation-driven design and explain how CAESES, our parametric geometry exploration and optimization platform, helps teams overcome them.

1. Creating Robust Parametric Geometry Models

One of the biggest challenges in simulation-driven engineering is building robust parametric geometry models. Most CAD tools are focused on detailed design and manufacturing intent. Their models only function within narrow “safe” parameter ranges. As soon as parameters move outside these ranges, the geometry fails to regenerate or produces unreasonable shapes.

Typical causes include:

• Fragile feature-based CAD histories
• Unstable topology changes
• Models not designed for automation or optimization

Why this matters to engineering teams:

Broken geometry interrupts automated workflows, delays simulations, and forces engineers into time-consuming manual fixes.

How CAESES helps:

CAESES is designed specifically for robust parametric modeling for shape variation. Its dependency-based surface modeling approach ensures geometry remains stable across wide parameter ranges, making it ideal for design space exploration, CFD automation, and optimization studies.

2. Identifying Dominant Design Parameters

View full image

Modern engineering models often include dozens of parameters, but not all of them influence performance in meaningful ways. Teams frequently struggle to identify which design variables actually matter.

Common issues include:

• Too many parameters with unclear impact and redundancies
• Limited visibility into sensitivities
• Time wasted optimizing low-impact variables

Why this matters:

Without understanding dominant design drivers, optimization efforts become inefficient and unfocused.

How CAESES helps:

CAESES supports design sensitivity analysis and structured design exploration, helping engineers quickly identify the parameters that drive performance. This allows teams to reduce complexity, focus optimization efforts, and make data-driven design decisions.

3. High-Dimensional Design Spaces with Competing Objectives

View full image

Real-world engineering problems rarely involve a single objective. Teams must typically cover vast design spaces and balance performance for different measures of interest and/or operating points, often within limited development timelines.

As design complexity increases:

• Manual iteration becomes impractical
• Trade-offs are difficult to quantify
• Good designs are hard to uncover, and teams settle for “good enough” designs

Why this matters:

Without systematic optimization, teams risk missing better design solutions.

How CAESES helps:

CAESES enables automated multi-objective optimization, allowing engineers to explore high-dimensional design spaces efficiently. By visualizing trade-offs and Pareto fronts, teams can confidently select optimal design compromises based on data and not intuition.

4. Tightly Constrained Optimization Problems

Many engineering projects are governed by strict constraints such as:

• Packaging and installation limits
• Manufacturing rules
• Regulatory or performance thresholds

In such cases, most randomly generated designs are infeasible, causing optimization workflows to fail or slow down.

Why this matters:

Optimization algorithms waste time evaluating invalid designs, and engineers spend effort fixing geometry instead of improving performance.

How CAESES helps:

CAESES integrates constraints directly into the parametric geometry definition, ensuring that only feasible designs are generated. This dramatically improves optimization efficiency especially for tightly constrained engineering problems.

5. Schedule Overruns Due to Manual Redesign Loops

View full image

Late design changes are inevitable but manual redesign loops shouldn’t be. Many schedule overruns occur because:

• Requirements aren’t fully integrated early
• Geometry models lack flexibility
• Changes cascade across CAD, meshing, and simulation

Why this matters:

Each redesign loop increases cost, risk, and time-to-market.

How CAESES helps:

CAESES enables requirement-driven parametric design, allowing geometry to adapt automatically as inputs change. This reduces manual rework, keeps simulation workflows consistent, and helps engineering teams stay on schedule.

Conclusion: Enabling Smarter Engineering

View full image

The challenges engineering teams face today are not caused by a lack of tools but by inefficient geometry and workflow management.

By combining robust parametric modeling, design automation, and optimization-ready geometry, CAESES helps teams:

• Build stable parametric models
• Identify key performance drivers
• Explore complex design trade-offs
• Solve tightly constrained optimization problems
• Reduce costly redesign cycles

CAESES empowers engineering teams to move from trial-and-error to data-driven, simulation-first design without sacrificing speed or reliability.

With CAESES 5.4, FRIENDSHIP SYSTEMS delivers the next step in the evolution of CAESES, further expanding the maritime design functionality introduced with CAESES 5.3 while also introducing powerful new features that benefit all CAESES users. The new release provides even more flexibility and support for naval architects, as well as improvements in geometry modeling, parameterization, and computational workflows.

Extended Component-Based Ship Modeling Capabilities

CAESES 5.4 continues to build on the component-based ship hull modeling workflow, further strengthening CAESES as a central tool for parametric ship design and optimization. The extended set of components allows users to create complex and realistic ship geometries in a modular, structured, and fully parametric manner.

View full image

Newly added components for ship hull modeling include:

• Waterline-based aftship
• Section-based foreship
• Sweeping foreship
• A new F-Spline version of the waterline-based foreship
• Center skeg for twin-screw vessel configurations
• Rudder components, including full-spade, balanced, and generic rudder types
• Propeller components, including actuator disc and generic propeller representations

These additions further empower naval architects to efficiently model a wide variety of ship types and configurations while maintaining full geometric control and variability. The component-based approach is particularly well suited for rapid model generation, as well as following design space exploration and optimization workflows, where consistency and robustness are essential.

Dedicated Workflow for Hard-Chined and Planning Boat Hulls

In addition to displacement vessels, CAESES 5.4 introduces a new dedicated workflow for hard-chined (planning) boat hulls. The workflow is based on a bare hull component and can be extended with optional functional elements, including:

• Spray rails
• Propeller tunnels

View full image

This new workflow provides a clean and efficient setup for modeling high-speed craft and planning boats, while fully preserving the parametric and associative nature of CAESES models. It allows designers to rapidly evaluate variations and assess performance-relevant geometric changes.

Enhanced Floating Conditions for Hydrostatic Computations

Hydrostatic analyses in CAESES 5.4 have been extended with additional options for defining floating conditions. Users can now specify floating conditions based on:

• A given trim angle and vessel mass
• Draft marks, defined by two longitudinal positions and the corresponding drafts at those locations

View full image

These new options provide greater flexibility when analyzing vessel behavior under different loading and trim scenarios, and better reflect real-world use cases commonly encountered in naval architecture projects.

Parametric Sample Models for Benchmark Vessels and Propellers

To support validation, simulation, and optimization studies, CAESES 5.4 includes fully parametric sample models of two widely used maritime benchmark vessels:

• KRISO Container Ship (KCS)
• KRISO Very Large Crude Carrier (KVLCC2)

Both models are implemented using the component-based ship hull workflow and yield simulation-ready geometries suitable for meshing and CFD analysis. At the same time, the models provide full geometric variability, making them ideal for optimization studies and method development.

View full image

In addition, new sample models are provided for a wide range of propeller types, including the well-known Potsdam Propeller Test Case (PPTC), further supporting propeller analysis and validation workflows.

Improved Geometry Export for Maritime Applications

CAESES 5.4 introduces a new option to export trimmed surfaces using IGES entity 144, as an alternative to exporting BReps or assemblies. This option is particularly helpful when transferring ship hull geometries to downstream tools such as NAPA, ensuring a smoother and more robust data exchange.

Advanced Curve Parameterization for Improved Surface Quality

Several new curve parameterization options have been added in CAESES 5.4, giving users more flexible control over the speed of the curve parameter t∈[0,1] along a curve. These options make it possible to:

• Align the parameterization speed of multiple curves to a reference curve
• Achieve more consistent parameter distributions
• Create cleaner surfaces with better-ordered control polygons

View full image

These enhancements are especially beneficial for high-quality surface modeling and for ensuring robust downstream operations such as meshing and optimization.

Introduction of the CTC Server (CAESES Task Controller)

A major new infrastructure feature in CAESES 5.4 is the introduction of the CTC Server (CAESES Task Controller). The CTC Server is a modular system designed to extend CAESES with advanced job scheduling and task management capabilities.

View full image

In its first release stage, CAESES 5.4 includes the Slurm Bridge module, which enables CAESES to interact directly with SLURM (Simple Linux Utility for Resource Management). Through this integration, CAESES can:

• Submit jobs to SLURM queues
• Monitor job states
• Cancel or reschedule running jobs
• Provide job monitoring via a lightweight web interface

This functionality is a key enabler for more scalable, automated, and HPC-oriented simulation workflows.

Numerous Additional Enhancements and Usability Improvements

Beyond the major features, CAESES 5.4 includes many smaller but impactful improvements across geometry modeling, usability, and visualization, such as:

• Custom labels for operations, making complex object trees easier to navigate
• Surface curvature visualization while interactively moving points
• Automatic slider creation in the object editor when design variables with bounds are defined
• A new BRep operation to color edges or edge sets based on a reference point
• A new option to create fillet surfaces with fixed tangent lengths
• A new NURBS surface command to propagate knots from a neighboring surface by raising the degree and inserting matching knots
• A new BRep command to check whether a given point is located inside a manifold BRep

Getting Started with CAESES 5.4

CAESES 5.4 is now available and ready to support advanced, simulation-driven design workflows – from maritime applications to general-purpose parametric geometry modeling.

For further details, please consult the CAESES documentation, the full change log, and explore the new sample models and workflows included with the release.

The advantages of early-stage Design Space Exploration (DSE) are well recognized within the fast-ship design community. By evaluating design variations in the process, DSE can significantly influence both fundamental design and business decisions, and enhance the effectiveness of later simulations that rely on resource-intensive, higher-order methods.

This study is a collaborative effort between HydroComp Inc. and FRIENDSHIP SYSTEMS AG. It was presented at the SNAME FAST 2025 Conference, and serves as an extension of the AutoPlan R&D project, which included viscous CFD analyses, towing tank experiments, and full-scale sea trials.

In AutoPlan, the optimization of an 11-meter planing craft operating at 27.5 knots was performed using high-fidelity tools such as RANS-based CFD (Simcenter STAR-CCM+). While these methods provide highly detailed and accurate results, they are computationally expensive and time-consuming, making them less suitable during conceptual design stages.

This raises the central question:

Can faster, reduced-order simulation tools be used to effectively explore the design space with enough confidence to support informed design decisions?

To answer this, the study follows the same design process used in the AutoPlan project, allowing side-by-side comparisons of outcomes and resource requirements, and replaces the high-fidelity CFD simulations with a resource-efficient, reduced-order workflow using CAESES® for parametric hull design and NavCad® for early-stage performance prediction.

Methodology

The table below summarizes the methodology followed in this study for comparing the outcomes and resource requirements of the AutoPlan project:

FC-DSE

FS-DSE uses a simplified force model that excludes appendages and does not account for propeller-induced forces.

Computation is performed at model scale to reduce computational cost, and full-scale values are obtained using a nonstandard expansion method, in which power is treated as an exponential function of the scale ratio.

SERO-DSE

SERO-DSE uses a more complete force model that includes lift from appendages and propellers, as well as spray drag. Computation is performed at model scale using dimensional analysis of model test data, and expansion to full scale is carried out through a combined Froude and Reynolds scaling approach.

Obviously, both approaches represent simplified models of reality. Neither is fully correct, and each is an approximation that is useful in different ways. The distinction lies between the data model, which defines how the hull and forces are represented and applied, and the analytical model, which defines how the calculations themselves are performed.

Validation

Of course, speed is useless without accuracy, so the first step was validation. The NavCad predictions for the baseline hull were compared with the original model test data and CFD results. As shown in the drag curve, the predictions are in very close agreement, especially at the design speed, providing strong confidence in the workflow.

View full image

Data Flow

As shown below, the connection setup between CAESES and NavCad illustrates how data flow is managed and exchanged between the two software packages using batch scripting. Parametric modeling is carried out in CAESES, which triggers the simulation process, while NavCad performs the corresponding calculations and returns the resulting performance outcomes.

View full image

Parametric Model in CAESES

A parametric model, using the same parametrization employed in the AutoPlan project, was developed in CAESES. The animations below illustrate the design space and the range of design variables explored.

View full image

View full image

View full image

View full image

View full image

The NavCad Simulation

Regarding the NavCad simulation, it is useful to look briefly under the hood at how the system operates.

Resistance Prediction

Resistance is computed using an equilibrium method based on the well-known Savitsky (1964) planing theory, which has been enhanced with modern extensions that account for hull tunnels, spray drag, and bow-wave formation. A key component is the propulsor lift model, which captures the vertical forces generated by the shaft line described earlier.

Propulsion Simulation

Propulsion is handled through a more advanced process than a single, fixed calculation. For each hull variant, the simulation performs its own sub-optimization to determine the ideal propeller configuration for that specific design. This approach ensures a fair, like-for-like comparison of the best achievable performance for every hull. Performance is then evaluated using a steady-state equilibrium condition in which thrust and resistance are balanced.

View full image

The Optimization Process

Since the modeling phase, simulation phase, and connection setup are covered, an automated loop can be applied, enabling design space exploration and optimization driven by CAESES.

Design of Experiments (DoE)

A Sobol sequence was used to efficiently explore the design space and identify key relationships between hull parameters and performance.

Local Optimization

A T-search method was then used to refine the most promising variants identified in the DoE.

Comparable Performance Gains

A direct comparison is difficult because the CFD analysis was conducted at model scale, used a simplified appended-drag model and applied a scaling procedure that differed from the standard ITTC approach. For this reason, the comparison is based on relative improvements rather than absolute values. Both paths, however, resulted in significant and similar reductions in drag and power.

Comparing relative improvement

• FC-DSE (CFD): 10.9% reduction in shaft-line thrust, which correlates directly with total drag and an 11.6% reduction in shaft power
• SERO-DSE (NavCad): 11.9% reduction in drag and a 14.9% reduction in shaft power

The difference in shaft-power improvement relative to drag is attributed to propeller sizing effects and oblique-angle correction.

Comparing Baseline with SERO-DSE Optimized

Geometric Similarity

Even more compelling is that both methods produced remarkably similar final designs. The CFD-optimized hull is shown (see image below) in purple and the semi-empirical hull in green, while the ghosted blue line represents the baseline hull used for both studies. Both approaches converged on similar strategies for improving efficiency, such as narrowing the chine beam and modifying the deadrise. This demonstrates that the reduced-order tool is not only producing accurate performance estimates, but also identifying the same geometric trends.

• Chines shifted inward and downward
• Rising keel resulting in a flatter deadrise
• FC-DSE exhibits a higher static draft, driven by a narrower chine beam and increased bottom warp; the deadrise is higher forward and flatter aft

View full image

Gains in Workflow Efficiency

Putting everything together, the semi-empirical approach delivered performance gains comparable to those from CFD and produced an optimized geometry that is nearly identical. A full design-space exploration using the semi-empirical tool can be completed on a laptop in less time than it takes to run a single CFD variant on a computing cluster. This does not imply that CFD lacks value or should be ignored; rather, it suggests that CFD should be positioned later in the design process. By investing less time and fewer resources early on, fundamental design decisions can be made before committing to costly CFD analyses, ensuring that the resulting data is more relevant, more accurate, and more useful to the design team.

Conclusion and Future Work

In conclusion, this study demonstrates that semi-empirical, reduced-order simulation provides a powerful framework for early-stage marine design. It is not a replacement for CFD, but serves as a highly effective front-end filter that identifies the most promising candidates for detailed analysis. The next steps involve expanding the methodology to include off-design performance and seakeeping criteria.

Duty profile optimization

This effort will extend the focused design-point optimization to include performance at off-design speeds. By defining a time- or distance-based duty profile, the objective function can be formulated as total mission fuel consumption, accounting for partial-load efficiencies of internal-combustion engines and electric motors.

Observation of additional performance parameters

Although efficiency at speed is typically a key optimization metric, the optimum design may or may not satisfy other critical criteria. It is therefore important to incorporate additional performance parameters into the optimization process. These include longitudinal and transverse dynamic stability (such as porpoising) and seakeeping characteristics (such as impact accelerations in irregular seas).

Refinement of shape parameters

A new prediction model for stern lift associated with propeller-tunnel aft-exit curvature is currently under development.

The full paper that was presented at the FAST 2025 conference is available from SNAME here.

Traditional Drafting

For centuries, engineers began every concept with a sketch. A few lines on paper were enough to structure ideas, estimate proportions, and visualize an outcome before anything was built. Before computers, designs were drawn by hand using traditional instruments such as rulers, compasses, and protractors. Yet freeform shapes, such as a ship’s bow, could not be captured precisely with these rigid tools.

To draw smooth curves at full scale, engineers used flexible wooden strips known as splines, held in place by lead weights called ducks. Between these anchor points, the spline naturally formed the smoothest possible curve, achieving both precision and elegance. The drawing board served not just as a workspace but as a physical extension of an engineer’s thought process.

View full image

Computer-Aided Design (CAD) Modeling

As computers entered the design process, engineers began to study the physical properties of splines so they could be modeled with mathematical precision and reproduced digitally whenever needed. The iterative definition of basis functions in NURBS (Non-Uniform Rational B-Splines) aligned perfectly with the increasing computational power of early computer systems, enabling levels of computational calculation, precision, and B-Rep (Boundary Representation) modeling far beyond human capability. The transition from mainframe computers to minicomputers and eventually to personal computers in the 1980s made computer-aided design (CAD) accessible to a much wider audience. The mouse replaced the pencil, becoming a new extension of the human mind. It offered enhanced precision, digital standards, and file formats that streamlined workflows and reduced manual errors.

While CAD revolutionized efficiency, it did not fundamentally change the underlying approach of design. Similar to traditional drafting, CAD still relied on the designer’s judgment to ensure overall consistency by placing geometric primitives or digital annotations within a virtual drawing space. The concept of layers in CAD reflected the additive logic of the traditional drawing board, now translated into the digital realm.

View full image

Parametric Modeling and Algorithmic-Aided Design (AAD)

In the 1990s, major advancements in 3D modeling and parametric design introduced feature-based solid modeling. These developments significantly changed how designers and engineers approached geometry, giving rise to modern parametric and algorithmic-aided design (AAD).

Algorithms began to generate geometries directly. When an integrated editor was used within CAD or other modeling software, objects were no longer manipulated by hand or with a mouse. Instead, they were defined through procedures expressed in programming languages.

This approach, known as scripting, was relatively new to design practice. It redefined the connection between the designer’s idea and the final output. The result was no longer a static digital drawing but an interactive digital model that could respond dynamically to changes in input, manipulating the entire system accordingly.

The combination of human judgment, computer programming, and engineering methods expanded creative and analytical capabilities beyond what could be achieved manually.

View full image

Simulation-Driven Design (SDD) with CAESES

CAESES®, introduced in 2001, provides a specialized CAD environment for parametric modeling and integration with simulation tools. It includes its own scripting language, the Feature Programming Language (FPL), which allows users to automate modeling tasks, create custom features, and link external tools directly to the design process.

Anything that can be executed in batch mode can be connected to CAESES, making it particularly effective for coupling with CFD codes and other simulation software. It also offers a complete optimization framework for exploring and refining the design space through Design of Experiments (DoE) and advanced optimization algorithms.

This process defines simulation-driven design (SDD), where simulations guide engineers toward optimal solutions based on data and performance rather than intuition. For the past 25 years, simulation-driven design has led the way in advancing modern engineering.

View full image

Future with Data-Driven Engineering (DDE)

The future is increasingly shaped by the rise of generative AI. Artificial intelligence has long been familiar to engineers through methods such as linear regression, machine learning, and surrogate modeling, but in recent years, the rapid growth of generative AI and large language models (LLMs) has shown that AI-based systems and intelligent agents are becoming an essential part of engineering.

The focus is gradually shifting from simulation-driven design (SDD) toward data-driven engineering (DDE). Engineers today often possess extensive simulation data that requires significant computational time to produce. It is not efficient to rerun new simulations for every design variation, especially when these computations can take hours or even days and require access to high-performance computing systems or large clusters.

When data is already available, the goal is to develop a machine learning-based surrogate model that can instantly retrieve results, even on a standard laptop. When data is limited, engineers can use Design of Experiments (DoE) algorithms, such as a Sobol sequence, to explore the design space efficiently, performing as few simulations as possible to build accurate surrogate models. These models then enable quick predictions and design evaluations without the need for repeated costly simulations.

This approach also enhances collaboration among colleagues. Once the setup is complete, users do not need to be CFD experts to explore design alternatives. With a well-prepared parametric model, they can adjust parameters and immediately observe how performance metrics change.

However, one must always consider the limits of extrapolation. The accuracy of machine learning and surrogate models depends on the quality of the data, the coverage of the design space, and the nature of the underlying physical problem.

Therefore, a human expert remains essential to evaluate and validate the outcomes, and across all stages of design and manufacturing, the process remains human-centric. The engineer continues to think, decide, and create, while the medium evolves to meet new challenges. Each new generation of tools, from pencil to code to generative AI, expands the creative and analytical power of the human mind, enabling design with greater intelligence, creativity, and understanding.

View full image