Study on stress prediction model of EH36 steel in polar environments: Optimization of basis functions using adaptive genetic algorithm and simulated annealing

Abstract

As global temperatures rise, the Arctic route—a vital maritime corridor connecting Asia, Europe, and North America—is increasingly becoming a focus of international shipping. However, the extreme low-temperature environment of the Arctic presents significant challenges to the mechanical properties of shipbuilding steel, especially in marine engineering and ship design. This study investigates the mechanical behavior of polar-grade EH36 shipbuilding steel across a temperature range from -40 °C to 20 °C and strain rates from 0.00037/s to 5000/s, introducing an innovative approach for accurately predicting stress under these conditions. Comprehensive experimental data were obtained through a combination of quasi-static, low strain rate tensile tests, and high strain rate Split Hopkinson Pressure Bar tests. A stress prediction model was developed by integrating adaptive genetic algorithms and simulated annealing to optimize basis functions, enabling precise predictions across a wide spectrum of temperatures and strain rates. Model validation results demonstrate that the prediction error remains within 6 % under moderate strain rates from 1/s to 200/s, highlighting the model's high accuracy and broad applicability. This model not only overcomes the limitations of traditional experimental and interpolation methods but also provides essential data for the design and material selection of polar icebreaking ships, offering an efficient tool for engineers in extreme environments. These findings contribute to enhancing the performance and safety of vessels operating in polar regions.