Qualification of digital twins for predicting failure in welded structures

Abstract

Significant cost savings in Asset Integrity Management (AIM) are achieved by replacing condition-based with predictive-based maintenance strategies. However, this requires the gathering and analysis of huge amounts of real-time data from Structural Health Monitoring (SHM) sensors and periodic Non-Destructive Testing (NDT). This presents a problem that lends itself to Digital Twin (DT) Technology. In the case of process plant, DT technology has already been implemented using “Big-pipe” data from Condition Monitoring (CM) sensors to predict failures in pumps, compressors and other machinery on a company-wide scale. In the case of welded structures however, for example pressure vessels, pipelines and storage tanks, implementation of DT technology is hampered by a lack of confidence in the use of “Smallpipe” data from periodic NDT and continuous SHM. In “Big-pipe” data, data streams are very rapid flowing, with information directly available from the data. In “Small-pipe” data, data streams are very slow, accumulated over long periods of time and require careful interpretation to provide information for Engineering Critical Analysis (ECA) that is essential for predicting failure in the asset. This presentation will describe TWI’s approach to developing a DT for predicting failure in welded structures.