We spoke with Mirko Bromberger, Marketing Director at software provider Altair about what machine learning and AI can bring to businesses, especially for sustainability purposes. The company will exhibit at Hannover Messe some of their solutions
What are you showcasing this year at Hannover Messe?
Mirko Bromberger: “At this year’s Hannover Messe, we are highlighting simulation-driven design and AI-powered engineering as key elements. We want to show how machine learning contributes to the development process and discuss the widespread application of the digital twin. Digital twin still lacks comprehensive adoption across many industries. But we want to demonstrate how tasks can be expedited, and how it is possible to implement AI methodologies within organizations lacking AI experts. We provide solutions for domain experts. We want to empower them with a no-code platform featuring a visual interface, which eliminates the need for complex scripting or coding knowledge. This enables them to leverage the data from their machines or processes effectively.”
What can AI help businesses do?
Mirko Bromberger: “Our primary objective at Hannover Messe is to address the demand for AI implementation among clients who may not fully grasp its potential benefits. In terms of product development, AI offers diverse applications. For instance, in CAD modeling or simulation tasks, AI can help in geometry recognition, or expedite the model-building process by identifying similar parts. Thanks to multidisciplinary design exploration utilizing regression methods, AI refines results with greater precision.
Accelerating prediction times is another area where AI shines. This is the case for example in scenarios where traditional simulation runs are time-consuming. By leveraging geometric deep learning and feeding existing results into neural networks, companies can make predictions without additional simulations. This streamlines processes.”
Can you tell us more about romAI?
Mirko Bromberger: “In complex systems like thermodynamics or fluid-structure interaction, where real-time information is crucial, traditional co-simulation methods may prove too slow. However, by employing neural networks to process results from such intricate processes, and integrating them into larger system simulations, AI facilitates faster and more efficient analysis, even in control loop scenarios such as machinery operations. This approach, utilizing Reduced Order Models (romAI), enables dynamic or static behavior representation within control loops, enhancing system performance.
Another example illustrates the process of identifying the correct behavior. In the development of a mega-casting structure, various factors such as crash requirements, acoustics, safety considerations, and manufacturing constraints must be meticulously addressed. This multidisciplinary approach results in many individual scenarios, often exceeding 100, each requiring detailed examination. Tasks include topology optimization to optimize material placement, engineering redesigns for manufacturability, adjustments for acoustic properties, weight reduction, and crash safety enhancements.
When dealing with a large volume of simulations, clustering them becomes challenging. Integrating machine learning proves valuable in this scenario. It guides engineers in selecting parameters that align with desired outcomes. Through this iterative approach, engineers can identify key parameters influencing behavior and refine designs accordingly, paving the way for efficient and effective product development.”
romAI: How to speed up DEM simulations leveraging Artificial Intelligence (Credit: Altair)
Do we still require a dedicated AI specialist to manage this?
Mirko Bromberger: “There’s no requirement for a dedicated AI specialist to oversee this process. We’ve integrated it directly into the tool’s functionality. You can continue using your familiar CAD tool in the usual manner. And with the addition of this new wizard-like feature, you can expediently discover the optimal solutions.”
Do you intend to integrate co-pilots using generative AI?
Mirko Bromberger: “It’s less about having a co-pilot and more about providing an enabling environment to fully utilize the organization’s available data. We’ve developed a dedicated platform that spans across departments, from finance to procurement to production, housing various methodological approaches under one roof, with a user-friendly graphical interface for easy navigation.
As a domain expert, whether in engineering, finance, or science, you have the tools at your disposal to harness the power of AI for your specific needs.
Take material science, for instance. With our platform, even without conducting a test run, you can draw insights and accurately predict the performance of new material mixtures based on existing data.
Similarly, when it comes to recycling, you can use the platform to streamline ingredients and achieve the desired behavior.”
Can you give some examples of what AI can do to enhance sustainability?
Mirko Bromberger: “There’s a significant emphasis on energy conservation and sustainability within industries, aiming to reduce material usage and explore innovative alternatives. However, this often requires intricate engineering solutions due to the complexities involved in material substitution and geometric modifications.
For instance, PM Grupo from Italy specializes in manufacturing stamping equipment and sheet metal forming processes. They wanted to minimize production waste caused by varying parameters, such as temperature fluctuations affecting steel coils and changes in process requirements throughout the year.
To address this challenge, they implemented a system model integrating field data from sensors with simulation techniques, including romAI, to simulate diverse process scenarios. By feeding this data into a neural network, they developed a self-aware model capable of providing real-time feedback, enabling proactive adjustments to the manufacturing process. This approach resulted in a 15% reduction in waste.
Another example is Rupo Chim Bali, a manufacturer of professional coffee machines. Their substantial power consumption was approximately 6000 watts and they wanted to better manage their energy usage. To address this, they developed a digital twin model representing the machine and its energy dynamics across various operational scenarios. For example, the coffee machine may be idle in a coffee bar and suddenly experience a surge in demand when a large group arrives. By analyzing these scenarios and fine-tuning parameters, they optimized the machine’s performance to minimize energy consumption. This approach led to a notable 25% reduction in energy loss.”
Altair will be present at Hannover Messe Hall 17, Stand D25
RELATED ARTICLE
