High Cycle Fatigue Performance Of Lockbolt Ml40cr Steel Under Different Stress Ratios

Abstract

The high-strength lockbolt is a critical connecting system for wind turbines, providing exceptional anti-loosening performance compared with the traditional thread connecting system. ML40Cr steel, commonly used for lockbolts in China, offers excellent mechanical properties, but its fatigue behaviour remains underexplored. Therefore, this study investigates the high cycle fatigue (HCF) behaviour of ML40Cr steel, focusing on the factors that influence its HCF performance. These include external factors such as stress amplitude and stress ratio, as well as internal factors such as yield strength, ultimate strength and chemical compositions of the material. Fatigue tests were conducted under varying stress ratios, and the test results were compared with literature data on various ML40Cr and 40Cr steels, revealing that HCF life depends on both the stress ratio and the steel material itself. Furthermore, S-N curves from design codes were employed to evaluate fatigue life, showing that BS7608, GB50017-2017, and ANSI/AISC360-10 provided conservative predictions, whilst DNV and Eurocode 3 overestimated higher fatigue lives under specific stress ranges. Subsequently, the existing models, including Goodman, Gerber, Dietman, and SWT, along with a proposed model, were adopted to evaluate the HCF life of ML40Cr and 40Cr steels. They performed adequately in predicting the HCF life based on the equivalent stress amplitude by considering the stress ratio effect, however, failed to account for the influence of the steel material itself. To address this gap, machine learning models were introduced. These models achieved superior performance and accurately reflected the dependence of HCF life on both external and internal factors.

Keywords: High-strength lockbolt, High cycle fatigue, ML40Cr, Stress ratio, Fractography, machine learning