A digital twin-based method for predicting the burst speed of discs

Overspeed burst is a critical failure mode of discs, for which experimental validation is costly and destructive, while conventional finite element analysis (FEA) is computationally expensive and often inaccurate. To address these challenges, this study proposes a digital twin-based method for disc burst speed prediction by integrating multi-source uncertainty modelling, surrogate-based prediction, and adaptive model correction. A fracture strain criterion combined with FEA is used to estimate critical burst speeds, and a high-precision Kriging model is constructed via sequential sampling to improve prediction efficiency. Experimental data are incorporated to enable online correction of geometry and model parameters. Numerical results demonstrate that the proposed method improves prediction accuracy by 16.99% compared with conventional FEA, with the relative error of the actual burst speed reduced to 3.26%, while the computational time is reduced by approximately two orders of magnitude. The developed digital twin framework further supports the real-time updating of strain fields and the dynamic visualisation of the disc burst process. This method provides an accurate and efficient solution for disc burst prediction, supporting disc safety design and reliability assessment.