Campuses:

Enabling High-Fidelity Digital Twins of Critical Assets via Reduced Order Modeling

Tuesday, March 6, 2018 - 3:30pm - 4:30pm
Lind 305
David Knezevic (Akselos)
At Akselos we provide high-fidelity physics-based Digital Twins of critical assets from a range of heavy industry applications, such as offshore structures for oil & gas, wind farms, steam & gas turbines, compressors, and mining machinery. The key features that these systems have in common are: they are large and/or complex; they are critical in the sense that failures lead to safety and environmental risks as well as costly downtime and repairs; their reliability is driven by structural integrity issues such as structural fatigue, buckling, and cracking; and there is a need for rapid analysis either to perform re-analysis after an update based on new inspection or sensor data, to assess many what if scenarios, or to rapidly provide guidance on the best response to an accident. In this context it would be natural to apply high-fidelity structural simulation, such as finite element analysis (FEA) --- however, FEA is generally not capable of providing the speed and scale required for Digital Twins of large-scale systems. As a result, at Akselos we provide a scalable cloud-based simulation platform that is based on a novel reduced order modeling approach known as Reduced Basis FEA, or RB-FEA. RB-FEA builds upon FEA to provide fast and accurate reduced order models that scale efficiently to large-scale systems, and which provide a key enabler for high-fidelity Digital Twins.