(155b) Benefits of Combining Digital Twins with AI/ML

Anyone who has watched the movie Apollo 13 witnessed one of the most successful execution of the term “Digital twin” we hear so often these days. That happened in April of 1970. What was unique about this event was the manner in which NASA mission controllers were able to adapt and modify the complex simulations that the astronauts trained on the ground to match the conditions of the crippled spacecraft and bring the astronauts back home safely.

With the dramatic advancement in modern communications, compared to those that existed over 50 years ago, ‘Digital Twin’ paired with advancements in AI and ML (Deep Learning) is an increasingly valuable tool for transforming a wide range of industrial operations to a ‘high fidelity’ profile. The digital twin, provides ‘real-time’ profile of the asset or process of an entire plant or part of the critical operations of a plant. It provides insights on performance across all aspects of the process – operations, reliability, environmental performance, and maintenance, to name a few life-cycle performance items.

By pairing a digital twin with AI/ML, insights through predictions and rule-based prescriptive advice are provided. Performance optimization potential is determined and the changes needed to reach the optimized state can be defined. And, in all cases, the science (i.e. the digital twin) explains the data. However instead of having a descriptive tool that must be physically manipulated for diagnostic and predictive purposes, real time prescriptive and predictive advice is being provided at the operations level.

The challenges of combining a ‘high-fidelity’ Digital Twin of an asset with AI/ML and successful use-cases will be presented.