The Data-Driven Digital Twin: A Framework and Method for Transforming Operational Dredger Data into Improved Diagnostics
Type:
Presented during:
WODCON XXIII - Dredging is changing - The Practice. The Science. The Business.
Authors:
B.B. Visser, J. Osnabrugge
Abstract
" The dredging industry has started collecting large amounts of data for the purpose of answering questions such as: how can my dredging operations be more efficient or how do I reduce downtime of my vessel and equipment. But when the number of Inputs/Outputs (I/O's), the size of the fleet, the uniqueness of a vessel and the complexity of its various underlying systems begin to grow, managing all that data effectively becomes a challenge. This paper describes how this task can be split into the various components of the digital framework and proposes a digital twin-based approach that allows for the use of Deep Learning and Artificial Intelligence to effectively process large amounts of data. The Digital Twin offers an estimate of expected behavior under normal conditions and compares this with the real-life behavior of a vessel in order to recognize and identify system failure. This enables faster diagnostics, but also improves insight into the equipment with of the aim of optimization. The effectiveness and workings of this framework and approach will be clarified through the use of various examples taken from the dredgers' data collection efforts of Royal IHC."
Keywords: digital twin, deep learning, artificial intelligence, data framework, diagnostics