Time indexes in energy system optimisation: The example of oemof.solph

Abstract

In energy system modelling, time is typically discretised, i.e. handled in the form of time steps. While the choice of a time grid is a model parameter, it can affect the results of a simulation or optimisation to the same extent as physical parameters. This fact underlines the importance of design decisions by modelling frameworks, as these decisions define the options for modellers.

In oemof.solph, a system is implemented that allows a user-defined length for each single time-step. In the present contribution, we cover a detailed description of the status quo of the implementation, including two additional time index features, namely time-series aggregation and pathway planning, along with their experimental implementations. We then discuss options for harmonising the implementations, making them both, more user-friendly and easier to maintain. We also explain the implications of possible design choices on the capability to model the specifics of energy systems and provide examples for modelling and data-analysis with a focus on the use of time indexes.