Use of digital twins for process safety management

Abstract

The chapter discusses the framework for developing a Digital Twin (DT) for the process safety management (PSM) of small-bore piping (SBP) on a typical offshore platform. One of the important problems during the PSM of SBP is that due to significantly large number of SBPs on a process facility it is very difficult to place sensors at small bore connection (SBC) for stress estimation. In absence of the stress values, it is difficult to estimate the remaining fatigue life (RFL) of SBC which further impedes the inspection planning. Thus, in this chapter, a methodology comprising of CFD, FEA and Machine Learning is sued to obtain a virtual sensor for stress estimation at the SBC. The input to the virtual sensor is process parameters such as pressure and flow velocity while output is the maximum Von-Mises stress at the SBC. Thereafter, probabilistic crack growth law coupled with Bayesian Network is used to develop a DT for RFL estimation of SBP, which in turn is used to obtain reliability curves and inspection plans. Online deployment of the developed DT will give an up-to-date RFL estimates and inspection plans which can be used then be used for PSM of the SBP.