Using strain dynamics for fracture warning of shaft lining

Understanding the dynamic of shaft strain is of great importance for fracture warning of shaft lining. This study detects the statistical pattern of strain series during the fracture process of shaft lining by using the Rescaled Range Analysis (R/S) method, and two types of algorithms are employed for the comparison purpose to estimate the Hurst exponent. It is shown that the dynamical behavior of shaft lining system is a biased random walk in the long-term scale with Hurst exponent close to 0.6. This paper initiates the direction for reconsidering the fracture warnings of shaft lining based on the behavior of relative strain series. In addition, the rolling time window approach for the short scale analysis is adopted in this work to understand the time-varying correlation of cumulative strains and relative strains, which gives some deep insights into the fracture predictions.