Abstract
Abstract Real-time monitoring of the existing natural gas infrastructure using multiple sensors is crucial to achieving higher confidence and lower risk during hydrogen transportation. One of the main challenges during hydrogen transportation is that hydrogen can accelerate the pipe material's fatigue crack growth rate, which can be affected by operational variables like hydrogen gas pressure, load ratio, and load cycle frequency. The proposed solution will involve the development of a digital twin framework incorporating a data-driven fatigue crack model of existing natural gas pipelines. The digital twin system comprises three main components. A computational finite element-based model of the crack defect in the pipe is developed to calculate the stress intensity factor (SIF) for different crack lengths and depths. The computation model is used to estimate the fatigue damage. The input and output data from the computational model are used to develop a metamodel or surrogate model. The metamodel is a data-driven model typically used to replace the computational FE model in real-time fatigue damage monitoring or to estimate the remaining fatigue life. The data-driven and computational model outputs the fatigue damage or remaining fatigue life, which is an essential part of the decision-making process in the digital twin framework. Machine learning-based algorithms: Extreme Gradient Boosting (XGBoost) is used to estimate the SIF through a surrogate model. The result from surrogate model is compared with the Physics-based/ simulation model. This work uses Paris's law in the crack propagation model to calculate the fatigue crack's growth rate. Conventional S-N curve methods can only provide information about fatigue life, which typically corresponds to the point of fatigue damage initiation. In a digital twin system, the crack model predicts the crack growth or damage propagation which is a substantial advantage for real-time pipeline integrity management during hydrogen transportation. The proposed concept will provide a predictive early identification methodology for possible hazardous conditions specific to natural gas pipelines for hydrogen transportation.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have