Condition-based fatigue life monitoring of a high-pressure hydrogen storage vessel using a reduced basis digital twin

For the condition-based maintenance of high-pressure hydrogen storage vessels, we propose a statistical fatigue life monitoring scheme hinging on a digital twin realized by reduced basis analysis. The proposed strategy comprises four steps: inverse state estimation by Bayesian inference, equivalent stress evaluation by forward analysis, fatigue life prediction by the stress-life approach, and condition-based maintenance by statistical analysis. To expedite the entire monitoring process, we capitalize on reduced basis approximation that enables a rapid yet accurate structural simulation. For demonstration, we consider an intact pressure vessel subject to different internal pressures, followed by a damaged pressure vessel with a progressively increasing dent size over time. Numerical results confirm that the constructed reduced basis digital twins are nearly identical to the physical counterparts, as the sample means of estimated states deviate from the true states by less than 1%. Furthermore, statistical fatigue life prognosis using reduced basis digital twins is more expeditious than that using finite element digital twins by a factor of 52.86 on average. In conclusion, our proposed approach leveraging a reduced basis digital twin presents an effective solution for monitoring the fatigue life of pressure vessels, offering both accuracy and efficiency.