Due to the expansive, time-consuming nature of risk analyses, it is important to be able to assign the minimization of risk (and, in general, optimization of resilience) to responsible teams that can work in parallel. However, while methods exist for minimization of risk in conventional design processes, research has not yet shown how it should be performed in a model-based design context in early design phase, when the design representation is relatively high-level and there are uncertainties in parameter values. This paper presents a value-driven design approach to minimize risk by decomposing the design, operational, and expected failure costs to individual functions in a system failure model. This process is demonstrated in a case study considering the redundancy of components to fulfill overall functions in an electric power system, where it is shown to increase design value significantly. An uncertainty-based process is additionally provided to enable the designer to test the sensitivity of the chosen design solution to uncertain parameter values. In this limited case study it is shown that the sensitivity of the choice to parameter value uncertainty is low, provided the range of uncertainty for each parameter is within a reasonable range. In situations like this, presented expected cost metrics provide meaningful information to justify system-architectural design decisions made on the basis of resilience.