Efficiency of Simulations

With increasing product complexity, engineers face a consistent challenge of time and cost constraints, as the increasing number and complexity of the experiments leads to increasing time and costs. This has led engineers to trade time for the accuracy of their experiments’ results.

For decades now, simulations and 3D models have dominated engineering design. Simulation uses computational models to replicate real-world systems, enabling engineers to test, analyze, and optimize designs virtually, providing a cost-effective way to predict performance, optimize designs through rapid iteration, reduce costs by minimizing physical prototypes, and enhance safety without risking human lives. Simulations, even though time-consuming, offer the advantage of producing multiple results efficiently once the model is perfected.

This vast amount of results reminds me of visiting fine arts museums. A painter can create multiple versions of the same scene, each worthy of display. Conversely, a poet crafts a unique poem, singular and irreplaceable. Both are forms of art, but the medium of colour offers greater visibility. In the realm of computational fluid dynamics (CFD), sometimes humorously dubbed "colourful fluid dynamics," visibility is indeed key, making engineers artists. Don’t get me wrong. I understand how simulations can be an effective tool in engineering design processes. My question is ‘why?’

Embracing Surrogate Models

Let’s explore my question of the necessity of simulations, or to rephrase, why do we still need simulation. I’m not suggesting that engineers abandon their work for abstract explanations. The question then becomes, how can we make this entire simulation iteration faster? It is true that simulations allow rapid iteration and optimization, and significantly reducing the time-to-market for new products. But, the need to make better quality products while reducing design lead time is also increasing in a much faster pace.

With advancing technologies such as Artificial Intelligence (AI), we can now use surrogate models and reduced-order models (ROM) to optimize time and resources to address this challenge; by using a simplified version of the original model that retains key characteristics and behaviours. This reduction in complexity enables even quicker simulations, making it feasible to run more iterations in a shorter timeframe. We have leveraged this concept with our AI-powered surrogate modelling tool, nvision, that allows engineers to explore a broader range of design variations and scenarios efficiently, enabling them to predict their products’ performance in real-time, and conduct extensive parametric studies, that would be impractical with full-scale models.

Just as photography revolutionized art by opening new avenues rather than replacing painting, these models can enhance and extend the capabilities of simulations. nvision models, built from high-fidelity simulations or experimental data, provide a fast and efficient means to explore design spaces and make informed decisions without the computational expense of full simulations.

Then, do we still need simulations?

While surrogate models, built with tools like nvision, offer significant advantages in terms of speed and efficiency, they do not entirely replace the need for full-scale simulations. Using nvision models in conjunction with traditional simulations allows engineers, a balanced approach: high-fidelity models validating and refining the reduced models to ensure accuracy, while still benefitting from increased computational efficiency, to reduce cost and time. This synergy enhances the overall design and development process, pushing the boundaries of what can be achieved within shorter development cycles.

While nvision models can significantly speed up the iteration process and enhance efficiency, they should be built upon and validated by traditional simulations. As I always say, full-scale trustworthy simulation data provides the foundational accuracy, handle complex scenarios, and facilitate innovation, ensuring that the overall modelling and simulation ecosystem remains robust and reliable.

The Future of Simulations

As I continue to engage with the field of simulation, I will always encourage the use of surrogate modelling tools like nvision to save time and resources. The advent of AI and advanced modelling techniques doesn’t diminish the value of traditional simulations; instead, it opens new doors for creativity and innovation. By integrating AI and machine learning, we can develop intelligent systems that can predict outcomes and optimize designs more effectively than ever before.

While my background and interests may lean towards the experimental and the poetic side, I appreciate the critical role that simulations and AI-powered models play in modern engineering. Questioning and exploring the fundamentals is necessary to push the boundaries of what is possible, much like a photographer capturing new perspectives with each shot. Additionally, it facilitates the integration of simulation into real-time applications, such as digital twins and control systems, where speed is critical.

So, I ask again, why do we still need simulations? And I add, are we still scared of AI?

At Noesis Solutions, we are consistently evolving our product and service portfolio to adapt to the latest market and customer needs, and how our customers can leverage the use of AI/ML to gain competitive advantage. For more information, review the latest announcement, webinars, case studies and stay up to date on the latest products and case studies at Noesis Solutions. Watch a free demonstration of the product here and to discover how nvision can accelerate innovation at your organization, request a free non-obligatory consultation here.