The "design" year

Energy simulation programs don't refer to one, single year, except the case in which we use them to confront simulation to one specific but to a possible future and we need entered data to represent the weather patterns of a site. 
For this, climate experts, together with energy simulation experts, produce an artificial, or synthetic year that represents the most variation of outdoor temperature, solar radiation and the other energy-related based on a longterm average.
When we just want to compare design choices and their effectiveness and efficiency, we don't need real wether sequences but sequences able to represent average conditions as well as variations that change it far from the average, with the same frequency we may experience ; their correct choice is more important when we need to predict real conditions and performance outcomes, e.g. real energy savings and thermal comfort effectiveness. The most evident situations are urban case studies (weather dataset are based on data collected far from the city, explicitly to avoid the Urban Heat Island effect) and where the climate change is more effective.
Moreover, energy simulation results are very often used to assess the profitability of a design (new or retrofit) choice. So it may happen that we use, in 2018, a weather scenario developed from historical data older than 50 years, (e.g. from 50ies to 70ies) to predict 20 or more future years of energy consumptions and cost. Future is always  one of the less predictable thing, of course, and the uncertainty also depends on future energy costs, future maintenance costs and for very long prediction, on the efficiency of future plants (whose service life may be significantly shorter and end before the perspective date. Last but not least, not all the weather datasets show the same accuracy.
The question is not solved, at now. For example there is not a standard morphing procedure to revamp and rejuvenate 50 years old datasets. But researchers are studying the subject: see the wide literature review about data morphing in \cite{Cellura_2018} and as example, a case study, analized in \cite{Erba_2017}.