Dynamic modelling of variable renewable energy generation sources

Abstract

Renewable energy is one of the most important technologies for decarbonizing the economy and fighting climate change. In recent years, wind energy has become cheaper and more widely adopted. However, the variable nature of wind production creates unique challenges that are not faced by conventional thermal technologies. Several studies to date have showed the decrease in economic value of wind energy as penetration increases due to this variable nature. Plus, they also show that high wind penetration favors intermediate energy sources such as natural gas. I claim however, that few of these studies have considered the dynamic behavior and feedbacks of these systems, including investment delays and learning curves. This thesis uses system dynamics models to simulate the long term changes in the electric grid for Texas. The goal is to test two hypothesis: that the economic value of wind energy decreases as penetration increases, and that variable wind production favors natural gas technologies. It does this by calculating how wind energy changes the shape of the net load duration curve for a given region. This affect changes the profitability of different technologies in unique ways, due to their different fix and variable costs. The conclusions of this thesis are consistent with the literature, with the caveat that they are highly dependent on assumptions regarding the learning curve for energy technologies. The economic value of wind decreases, but this effect can be compensated by lower costs, leading to a continuing adaptation. Faster Wind adoption also reduces the profitably of technologies with high fixed costs such as coal and nuclear, and favors intermediate and peaking sources such as natural gas.