During last few years, researchers have been concentrating toward the estimation of stress–strain behavior of thin-film coatings using nanoindentation and finite element simulation. In present days, advancement of machine learning algorithms and artificial intelligence made it possible to implement a hybrid approach to extract the properties of a material in a more systematic way. The aim of this work is to find the optimum value of the constants of power-law plastic behavior of thin-film alloy coatings. In this study, an average load–displacement plot was obtained from nanoindentation tests of poly-alloy coatings. Then, a simulation data set was generated in ABAQUS according to design of experiment. A machine learning algorithm was used to generate the surrogate model correlating the constants of power-law plastic behavior as input and mean error between simulation and experimental results as output. Finally, optimization algorithm was used to find out the optimum values of constants of power-law plastic behavior.

在过去几年中，研究人员一直专注于使用纳米压痕和有限元模拟来估计薄膜涂层的应力-应变行为。如今，机器学习算法和人工智能的进步使得实现混合方法以更系统的方式提取材料特性成为可能。这项工作的目的是找到薄膜合金涂层的幂律塑性行为常数的最佳值。在这项研究中，从多合金涂层的纳米压痕测试中获得了平均载荷-位移图。然后，根据实验设计，在ABAQUS中生成了一个模拟数据集。使用机器学习算法生成代理模型，该模型将幂律塑性行为的常数作为输入，将模拟和实验结果之间的平均误差作为输出。最后，利用优化算法找出幂律塑性行为常数的最优值。