This paper focuses on the integrated design and control under uncertainties of a distributed energy system (DES): the future energy demand and production as well as the electricity tariff are not perfectly known while sizing and operating the system. To address this question, a two-stage stochastic programming model is implemented where uncertainties are modelled as random variables and the quality of the solutions is assessed on a common "out-of-sample'' simulator. The results show that the method provides designs which are consistent with a given risk measure (expectation, worst case, etc.).