Uncertainty Quantification for Decision-Making in Engineered Systems

摘要: This chapter discusses current research and opportunities for uncertainty quantification in performance prediction risk assessment of engineered systems. Model-based simulation becomes attractive systems that are too large complex full-scale testing. However, model-based involves many approximations assumptions, thus, confidence the result is an important consideration risk-informed decision-making. Sources both aleatory epistemic, stemming from natural variability, information uncertainty, modeling approximations. The draws on illustrative problems aerospace, mechanical, civil, environmental engineering disciplines to discuss (1) recent quantifying various types errors uncertainties, particularly focusing data model (both due form assumptions solution approximations); (2) framework integrating multiple sources (models, tests, experts), development activities (calibration, verification, validation), formats; (3) using decision-making throughout life cycle systems, such as design, operations, health assessment, management.