An enabling architecture for computational cost efficiency in predictive maintenance digital twins

Abstract

As digital twins emerge to provide a replication of physical assets in the digital space, the application of predictive maintenance of industrial asset becomes more effective. Developing digital twins for the predictive maintenance case study leverages Internet of Things, cloud computing and machine learning. While these technologies extend the necessary tools for deploying predictive maintenance digital twins, an enabling architecture facilitated by fog computing positions predictive maintenance digital twins for improved computational cost and latency than centralizing in the cloud or running locally at the edge. This work presents the application of a distributed digital framework, showing the benefits of better compute utilization and latency by adopting a distributed digital twin framework for predictive maintenance of wind turbine components in a wind farm.