Siemens Energy: How to decrease simulations computing days down to minutes.
Feb 21, 2024
auto_awesome
Siemens Energy AI expert discusses reducing simulation computing time with machine learning. Topics include turbine optimization, evolution of turbines in energy generation, speeding up engineering simulations, evolution of machine learning models, simulating material deformation, and efficiency with NVIDIA A100 GPUs.
Siemens Energy uses machine learning to reduce simulation time from days to minutes, enhancing turbine design efficiency.
Exploring generative design and 3D surrogate models, Siemens Energy aims for faster, more accurate turbine design predictions.
Deep dives
Application of Machine Learning in Industrial Design Process
The integration of machine learning in industrial design process is revolutionizing the approach to turbine optimization at Siemens Energy. By incorporating machine learning tools, the company aims to overcome computational limitations and enhance design robustness. Through a combination of commercial and in-house tools, like Heats AI Accelerator, simulations that once took days can now be completed in minutes, leading to increased efficiency in designing turbine components.
Advancements in Generative Design and Surrogate Models
Siemens Energy is exploring generative design and 3D surrogate models to predict and optimize turbine designs. By training models with data from various design scenarios, the company can swiftly analyze and optimize turbine responses across different environmental conditions. Utilizing machine learning techniques like random forests and Bayesian neural networks, Siemens Energy aims to achieve faster and more accurate design predictions, potentially reducing design efforts exponentially.
Automation of Creep Calculations with Machine Learning
Addressing the time-consuming nature of creep calculations, Siemens Energy is leveraging machine learning to expedite the process. By developing models that predict material behavior under different conditions, the company can accelerate the evaluation of turbine components subjected to varying temperatures and pressures. This automation allows for quicker analysis of complex structural changes, ensuring the integrity and performance of turbine components.
Future Outlook: Evolution of Industrial Foundation Models
Siemens Energy envisions a future where generative AI models collaborate with engineers to facilitate innovative design processes. While the industrial sector is in the proof-of-concept phase for utilizing machine learning in design, advancement towards a unified industrial foundation model is underway. The emergence of diverse models from startups and established companies suggests a shift towards creating comprehensive models that amalgamate simulation and engineering data for enhanced design precision.
Behnam Nouri is one of the AI experts at Siemens Energy. He talks to Peter Seeberg about the simulation of gas turbines and explains how he and his team were able to reduce the calculation time for simulations using machine learning.
Thanks for listening. We welcome suggestions for topics, criticism and a few stars on Apple, Spotify and Co.