Comparing time aggregation techniques in relation to capacity-expansion modeling

A high priority of green energy combined with a hard constraint of demand satisfaction causes system flexibility to be the core building block of a stable energy system. In order to both detect the need for flexibility as well as the optimal flexibility technologies satisfying these needs at minimum cost, the system should be analyzed on an hour-to-hour scale for a long period of time. This often leads to computationally intractable models and one way to regain tractability is
to aggregate the time domain. Many different aggregation techniques have been developed all with a common goal of selecting representative time slices to be used instead of the full time scale, causing a model size reduction by the number of variables and/or constraints. The art of aggregation is to balance the model complexity against the solution quality, making validation of the techniques crucial. We come up with a couple of new aggregation techniques, which we validate according to the full size model but also compare to other techniques from the literature. We look into the sensitivity of the performance of the techniques to different data sets and to different model features. With a focus on the complexity of the aggregation techniques, we try to answer the question whether more complex aggregations actually provide better estimates, and experimentally quantify the gains.