Data-driven conditional probability to predict fatigue properties of multi-principal element alloys (MPEAs)

Traditional fatigue assessment methods for new and unexplored metallic alloys is challenging due to very limited experimental data. To address this, we formulate the assessment within a conditional probability framework, allowing us to capture the complexities of uncertainty in fatigue predictions. We employ advanced probabilistic methods to account for both inherent material variability and model uncertainty. We analyse fatigue data of multi-principal element alloys (MPEAs) tested under stress ratios of R = 0.1 and R = −1, including face-centred cubic (FCC) microstructures of CoCrFeMnNi and AlCoCrFeMnNi alloys. Based on results, we found a clear material trend which allows us more reliable predictions and informed decision-making, which is distinct than the conventional fitting. for the future alloy design. As we demonstrate the efficacy of this framework in extracting the intricate relationships between MPEA composition and the trend in fatigue behaviour, we believe that our study will pave the way for enhanced advanced material design and uncertainty quantification in future MPEA research and materials engineering applications.