How AI-Powered Recommendation Systems Streamline CAE Workflows

Introduction

AI-Powered Recommendation Systems for CAE Workflows

In engineering, Artificial Intelligence (AI) tools are increasingly integrated into processes to support and relieve engineers. AI-powered recommendation systems are particularly transformative in Computer-Aided Engineering (CAE), streamlining workflows, enhancing efficiency, and opening new avenues for innovation. 

Despite their power, traditional CAE processes have limitations, such as repetitive tasks and manual effort. AI addresses these challenges by automating and optimizing various aspects of the engineering process. NAVASTO has been a pioneer in integrating AI into engineering, demonstrating significant enhancements in efficiency and accuracy through AI-enhanced tools.

AI recommendation systems are designed to suggest optimal actions based on vast amounts of data and complex algorithms. These systems comprise vital components such as data analysis engines, predictive models, and user-friendly interfaces in engineering. The NAVPACK API, developed by NAVASTO, exemplifies a flexible solution that integrates seamlessly into CAE workflows, providing real-time insights and recommendations.

Example: The Power of AI in Simulation and Design Optimization

Combining predictive AI with generative AI dramatically speeds up the design and simulation process. A Graph Neural Network predicts drag and lift for unseen geometries, which designers traditionally developed. 

The NAVASTO system rapidly produces inputs and results by replacing manual input with generative AI. AI-generated and predicted outcomes are available in seconds, transforming the traditional workflow into an optimization powerhouse.

Figure 1: Comparison between a manual workflow (left) and a workflow with an optimization loop (right). The manual workflow includes generating new inputs manually for AI predictions of unseen geometries. In the workflow with the optimization loop, unseen geometries are developed automatically by the generative AI while considering the given constraints.

 

The user sets the objective function and constraints, starts the optimization, and takes a break while the system iterates the design space to find the best, most optimum design option.

Figure 2: Visualization of the generated and analyzed AI recommendations for optimized drag and lift. The upper diagrams and the lower left show the number of generated and analyzed designs throughout 500 iterations until the optimum solution is achieved. The lower right diagram visualizes the parameter space with the drag and lift coefficients.

This workflow can be coupled to any optimization algorithm. A differential evolution with a subsequent local optimizer was used in this case. Still, the workflow allows us to switch to other approaches like surrogate-based optimization or adaptive Sampling.

The optimization loop substitutes the need for human effort to design single new inputs manually. Multiple geometries with matching goals are ready upon the user’s return from the coffee break. From robust machine learning models, these results offer the user several options:

• Select the best configuration and trust the AI prediction results.
• Choose promising configurations and run full high-fidelity simulations.
• Use suggestions for new product iteration directions.
• Adjust objectives and constraints, rerun optimization, and get new configurations.

This AI-driven process created many configurations in minutes, grounded in the existing design space. Constraints ensured customer-specific branding and avoided unrealistic shapes. Optimal for drag might resemble an egg, but practical designs needed nuance.Figure 3: AI Recommendation Process Output Example: A Car Geometry optimized for aerodynamic efficiency.

Navasto’s GenAI produces high-quality surface meshes that can be directly used for a prediction with the AI model but can also serve as an input for a simulation.

AI uncovered new connections in the design space, suggesting creative solutions. Some solutions might not be feasible but could inspire innovations. Ultimately, the engineer decided on the following steps, with AI providing valuable performance and improvement guidance.

Applications of AI Recommendation Systems in CAE

Several industry leaders, including General Motors, Volkswagen, and Formula 1 teams, have successfully implemented NAVASTO’s AI tools. These implementations have led to measurable improvements in workflow efficiency, demonstrating the tangible benefits of AI-powered recommendation systems in CAE.

Engineers use Navasto to preselect beneficial configurations. Since the generated geometries can also serve as input for simulations, they can choose whether they trust the prediction or want a high-fidelity simulation for the most promising candidates.

With NAVPACK, engineers receive tailored recommendations for simulation parameters, enhancing the accuracy and efficiency of simulations. NAVPACK’s automated design alternatives generation helps engineers explore multiple design options quickly, leading to more innovative and effective solutions. Furthermore, with NAVPACK and navDesign tools, real-time performance predictions are now possible, enabling engineers to foresee and address potential issues proactively.

Benefits of AI Recommendations in CAE

AI systems significantly enhance design optimization processes. For example, NAVPACK offers real-time insights that help engineers fine-tune designs efficiently, ensuring optimal performance and compliance with specifications. 

Additionally, tools like navDesign for ParaView and navDesign for Blender from NAVASTO enable immediate feedback loops, allowing engineers to make swift and informed decisions. This accelerates the overall engineering cycle and reduces the time to market. 

Leveraging AI’s predictive capabilities significantly reduces the frequency of errors and the number of iterative cycles. Case studies have shown that NAVASTO’s AI tools have substantially reduced errors in various industry applications. Furthermore, AI aids in current projects and in managing and utilizing accumulated knowledge from previous projects. NAVASTO’s solutions facilitate data-driven decisions and efficient knowledge management, ensuring valuable insights are retained and applied.

Considerations and Future Trends

Ensuring high-quality data is crucial for accurate AI predictions but remains a key challenge. Additionally, the possibility of seamlessly integrating AI tools like NAVPACK API is essential to minimize disruption and maximize efficiency. Building trust in AI is also vital, requiring reliable and transparent models that demonstrate consistent accuracy. 

Ongoing advancements in machine learning will enhance CAE workflows, increasing precision and efficiency. Finally, future integration with digital twins and IoT will create more sophisticated and interconnected CAE environments.

Conclusion

AI recommendation systems are revolutionizing CAE workflows, offering unprecedented efficiency and effectiveness. NAVASTO’s solutions exemplify AI’s transformative benefits in engineering, providing tools that enhance design optimization, decision-making, and overall workflow management. Engineers are encouraged to explore and adopt these cutting-edge technologies to stay ahead in the competitive engineering field.

By leveraging NAVASTO’s innovative AI tools, engineers can streamline their current workflows and pave the way for future advancements in CAE, ensuring sustained growth and innovation in their projects.

Visit NAVASTO’s website to learn more about how its AI solutions can transform your engineering workflows. For personalized assistance, contact the NAVASTO team.