Grid Connected Systems Final Project
In this project, I mimicked the process of writing an academic article through research, experimentation, and adhering to a strict writing style. I read a variety of journal articles in order to gather information, used NREL’s System Advisor Model (SAM), and created a model predictive control (MPC) algorithm in Python to run my tests.
Abstract
In this paper, a new time of use (TOU) pricing model is proposed and tested against existing TOU pricing models available to consumers in San Diego, California. The number of installations of photovoltaics (PVs) has increased hugely over a short period of time, creating a shift in the daily demand curve. What once was a relatively flat load profile now has a dip in the middle, which is causing a strain on infrastructure and an unintentional increase in carbon emissions due to fast-ramping peaker plants being used more frequently to supply electricity during peak load times. The solution in this paper implements a model predictive control (MPC) algorithm to optimize power draw from the grid while aiming to mitigate the effects of the duck curve and achieve a low daily energy consumption. The TOU-duck model resulted in the lowest daily energy consumption while also encouraging consumers to shift their power grid draw times in a way that helps to mitigate negative effects of the duck curve. The TOU-duck pricing model, like all TOU pricing models, relies on the assumption that not all people will opt into it. If that were the case, and if people shifted the times at which they draw power from the grid in order to save the most money, the demand curve would flip and the energy community would have the issue of a stegosaurus curve to deal with.
