Modern structures such as airplane wings and wind turbine blades exhibit a hierarchy of sub structures and typically make use of composite materials in their construction. Quantifying uncertainty in the strength and stiffness of composite structural materials is crucial for predicting the service lifetime of the structure. The high cost of experimental tests for large-scale hierarchical composite structures is driving a trend toward virtual testing. This requires the development of multi-scale numerical methods capable of handling large degrees of freedom spread across different length scales. In this talk we review model reduction strategies for multi-scale structural analysis in the presence of uncertainty as well as propose new multi-scale approaches that may be useful in predicting service lifetimes.