There is a growing demand for the development of precise, accurate, and industrially applicable predictive models for evaluating the fatigue performance of additively manufactured (AM) materials to accelerate their qualification. Six fatigue metrics based on microstructure-sensitive crystal plasticity finite element (CPFE) simulations are utilized to assess the fatigue performance of Inconel 718 produced via selective laser melting (an AM technique). Each metric is used to predict the probable locations of failure and the scatter in fatigue life under high cycle fatigue loading conditions. The predicted scatter from all the fatigue metrics was in good agreement with the experimental test data. The predictions locations of failure using two of the metrics, plastic strain accumulation, and plastic strain energy density, were found to correlate in a statistical sense with the post mortem fractography results. Moreover, an additional set of CPFE simulations were performed with varying volumes of the input microstructures, with the largest microstructure having a volume close to that of the test specimen's gauge section. This additional analysis was intended to provide informed guidelines for simulation volume that is both computationally tractable, and results in consistent scatter predictions, which was estimated to be a simulation volume consisting of ∼200 grains. Lastly, fatigue life predictions obtained from traction-free and more realistic microstructure constraints were compared to understand the role of boundary conditions (BCs) in fatigue life predictions. This work is beneficial in determining the statistical minimum for safe-life analyses, thereby reducing the overall number of experimental tests and accelerating the qualification process.

越来越需要开发精确，准确和工业适用的预测模型，以评估增材制造（AM）材料的疲劳性能以加速其鉴定。基于微观结构敏感的晶体塑性有限元（CPFE）模拟的六个疲劳指标可用于评估通过选择性激光熔化（AM技术）生产的Inconel 718的疲劳性能。每个度量标准都用于预测在高周疲劳载荷条件下疲劳寿命中的失效位置和散布。来自所有疲劳度量的预测散布与实验测试数据非常吻合。使用塑性应变累积和塑性应变能密度这两个指标来预测失效位置，被发现在统计学意义上与验尸后的断层扫描结果相关。此外，使用不同体积的输入微结构执行了另一组CPFE模拟，其中最大的微结构的体积接近于试样标尺截面的体积。这项额外的分析旨在为模拟量提供明智的指导，该模拟量在计算上是易处理的，并且会导致一致的散射预测，估计该散射量是由约200个晶粒组成的模拟量。最后，比较了从无牵引力和更现实的微观结构约束获得的疲劳寿命预测，以了解边界条件（BCs）在疲劳寿命预测中的作用。这项工作有助于确定安全寿命分析的统计最小值，