The adequacy of time-series reduction for renewable energy systems.

To manage computational complexity, models of macro-energy systems commonly deploy reduced sets of time-series data. This paper evaluates the adequacy of time-series reduction when modelling energy systems with fully renewable generation and a consequent dependency on storage. Analysis includes various methods to derive reduced time-series and to implement them into models, either as time-slices, also referred to as representative days, or continuous time-steps. All methods are tested with regard to unmet demand and accuracy of estimated system costs using a simple capacity expansion model of the power sector within a renewable energy system. Some methods achieve little unmet demand, but instead their results regarding storage are biased and favour seasonal at the expense of short-term storage. We conclude that renewable energy systems limit the adequacy of time-series reduction and future research should focus on alternative methods to reduce computational complexity.