Abstract :
[en] In the eve of a climate crisis generated by the sustained combustion of fossil fuels across various economic sectors, decarbonising worldwide power systems has been a cornerstone in reaching net-zero targets in the upcoming decades. To this end, widely-available renewable energy sources (RES) such as solar irradiance or wind have been recently harnessed at scale in order to replace fossil-based generators in the electricity mix of power systems around the world. However, such resources are inherently variable on time scales ranging from minutes to years and integrating them in power systems typically complicates planning and operational procedures. Several solutions have been advocated to alleviate these issues, including the large-scale deployment of electricity storage systems or the implementation of demand response programs. Alternatively, since RES are heterogeneously-distributed in space and time, it has been suggested that siting RES electricity production assets so as to exploit this diversity may reduce the aggregate output variability of power plants as well as the residual electricity load (i.e., total load minus renewable production). The concept of renewable sources spatiotemporal complementarity formalises this idea and makes for the chief concept investigated in this thesis.
The manuscript starts by revealing how connecting remote RES sites could lead to reduced probabilities of low-generation events. Then, a framework explicitly designed to assess the spatiotemporal complementarity between geographically dispersed RES assets is introduced and leveraged to devise optimisation models seeking to identify deployment patters with maximum complementarity among sites. Once an optimisation problem for siting RES assets based on complementarity criteria is made available, the value of spatiotemporal complementarity for power systems is assessed. Essentially, this is made possible via a multi-stage approach that works as follows. In the first stage, a highly-granular siting problem identifies a suitable set of sites where RES assets could be deployed according to a pre-specified criterion (e.g., spatiotemporal complementarity, output maximisation). In the second stage, the subset of previously identified sites is passed to a capacity expansion planning framework that sizes the power generation, transmission and storage assets that should be deployed and operated in order to satisfy pre-specified electricity demand levels at minimum cost. Furthermore, a third stage may also be leveraged should a more accurate estimation of the impact of different siting criteria on the operation of power systems is sought. This stage is formulated as a classical unit commitment and economic dispatch problem and, given the capacities of power generation, transmission and storage assets resulted from the second stage, provides a more detailed view on the daily operation of the power system assets. Finally, inspired by the workings of the aforementioned routine, a method to reduce the spatial dimension and decrease the computational burden of capacity expansion planning problems while preserving a detailed representation of RES assets is proposed.
