E8, AI Deep Learning-Based Real-Time Simulation Technology Development

Simulation-based digital twin platform company E8 announced on the 20th that it has developed a simulation model applying AI deep learning technology to computational fluid dynamics (CFD).

Computational fluid dynamics (CFD) is a methodology for numerically solving problems related to fluid flow. It can perform virtual simulations of various physical phenomena. However, as the scale of calculations increases, the time required for computation also increases, which has limited its application in fields requiring near real-time results, such as disaster and emergency management.

In response, E8 developed an AI-CFD deep learning model that drastically reduces the simulation time compared to conventional CFD. This model is a reduced-order model applying Beta-VAE (Variational Auto Encoder) technology to E8’s NFLOW. NFLOW generates simulation results, and based on the generated simulation data, the model is trained to produce analysis results that previously took over 30 minutes in near real-time.

The most critical factor in disaster simulation is real-time capability. In disaster situations, it is essential to immediately reflect emergency conditions to support rescue teams and emergency services in finding optimal routes and responding swiftly. Therefore, real-time data processing and analysis capabilities are core elements of disaster simulation, playing a vital role in effective disaster management and damage reduction.

E8 is participating as the digital twin builder for the Sejong Smart City, a national pilot city project, conducting disaster simulations such as wind corridors and floods. The newly developed AI-CFD technology will be applied, with plans to expand its application areas to various industries including engineering, energy, construction, and nuclear power.