Variability and variation management in a renewable electricity system -large-scale wind- and solar power deployment in Europe

Doctoral thesis, 2017

The large-scale deployment of wind and solar power poses challenges to the electricity system by introducing variability on the generation side. To handle these variations, variation management strategies can be employed to ensure that generation meets demand. Such strategies include the deployment of storage and investments in wind and solar capacities in different regions connected by a network of transmission lines. This thesis investigates the interplay of such strategies, with special focus on the geographic distribution of wind power in Europe. The research questions addressed here include: To what extent can geographic distribution of wind power tailor aggregate output? What are the main characteristics of a cost-effective electricity system that is based on renewable energy? How does the system cost increase with the penetration levels of wind power and solar power? How should the time dimension be handled in electricity investment models that aim to design systems with a large share of variable renewable energy (VRE)? Five optimization models, characterized by comparatively high temporal resolution, are developed to answer these questions. Two of the models apply multi-objective optimization, whereby Conditional Value-at-Risk is used as a measure of the variation in wind power output/residual demand. Two of the models are network models that specifically target the effect of optimizing the transmission network so as to make better use of wind and solar power capacities. This thesis shows that the diverse weather patterns in Europe can be exploited to achieve effective smoothing of wind power, provided that there is sufficient transmission capacity to trade variations in generation and demand. Furthermore, it is shown that dispersing wind power and building transmission lines is the most cost-efficient strategy to design an electricity system with a large share of renewable generation. A system that contains a large share of variable renewables is also shown to be versatile, in the sense that it is possible to combine large-scale variable generation with considerable amounts of base-load generation, provided that the transmission network is enhanced. The marginal cost for generation from variable resources increases in an approximately linear fashion with the penetration level of such resources, up to a VRE penetration level of around 80%. The marginal cost for achieving a VRE penetration level of 80% is approximately 50% higher than the initial cost. It is concluded that if wind- and solar power capacity is to be allocated over the entire area of Europe, at the same time as the transmission network is expanded, the cost for a future power system with a high (~80%) penetration of VRE will not be prohibitive.