重庆理工大学学报2025，Vol.39Issue(23)：186-193,8.DOI:10.3969/j.issn.1674-8425(z).2025.12.023
考虑参数随机性的有限元方法研究
Finite element method considering parameter randomness:a case study of indentation simulation
摘要
Abstract
Conventional finite element indentation simulation typically assumes one-dimensional or uniformly distributed material parameters,neglecting the inherent uncertainties present in actual experiments.This may lead to simulation results being inconsistent with test observations.The Box-Muller algorithm is employed to generate independent,multidimensional discretized samples of the mechanical properties of forged 18CrNiMo7-6 steel based on their mean values,standard deviations,and covariance structure,ensuring Gaussian-distributed inputs.The Sperman rank-correlation method is then adopted to embed the covariance relationships among the discretized material parameters.A total of 100 discretized indentation simulations are performed and their statistical reliability is evaluated by using a one-tailed F-test and a paired sample t-test.A confidence interval(CI)is provided for the reliability of the finite element simulation input parameters.For example,it provides a specified confidence level of 95%.Results show the distribution trends of the discretized data samples are in line with the tensile test results.Multiple sets of discretized simulations effectively capture the impact of uncertainties during the testing process on the experimental outcomes.Simulation outputs falling within the derived confidence intervals have a 95%probability of reflecting the true mechanical response of the material.关键词
微纳米压痕/有限元仿真/离散化/压痕形貌Key words
micro-nano indentation/finite element simulation/discretization/indentation morphology分类
机械制造引用本文复制引用
LI Hongtao,JIN Rong,JIANG Qinlu,LIU Haitao..考虑参数随机性的有限元方法研究[J].重庆理工大学学报,2025,39(23):186-193,8.基金项目
国家自然科学基金项目(U1804254,11702252) （U1804254,11702252）
河南省重大科技专项项目(201400211200) （201400211200）
