Presentation Information
[APP1-13]Strain-stress Coupling Cryogenic Test Platform Based on Digital Image Correlation
*LINjie Zhang1,2, Liancheng Xie1,2, Bixi Li1,2, ZIchun Huang1,2, Fuzhi Shen1, Hengcheng Zhang1, Hao Zhang1 (1. Key Laboratory of Cryogenic Science and Technology, TIPC, CAS (China), 2. University of Chinese Academy of Sciences (China))
Keywords:
DIC,PINN,Cryogenic,Strain-stress
In this paper, a two-dimensional-field strain-stress coupling cryogenic mechanics test platform is designed. Two-dimensional Green-Lagrange strain data is obtained by analyzing the displacement of the spot pattern which made by laser marking in specimen. The device system investigates system challenges such as extremely cryogenic temperature cooling or optical window design. A cooling model along the direction of heat conduction was established to analyze the cooling condition of the device and the results show that the temperature of the specimen can be adjusted and controlled at temperature in the range of 5~300 K. To obtain more accurate results, three methods are adopted to repair the result measured by DIC: Radial basis interpolation (RBF), Physics equation fitting optimization and Physics-informed Neural Networks (PINN). The full-field strain of 316LN stainless steel (SS) at room temperature (RT) and low temperature was measured using the DIC in the system. Evaluating the results repaired by the three methods, the RBF results are obviously defective. Compared with the Physics equation fitting optimization results, the average edge loss of the PINN result is reduced by 15%, and the average internal loss is reduced by 27%. Therefore, the results by PINN can be more accurate and reasonable than RBF and Physics equation fitting optimization, which verifies the reliability of the platform.