Abstract: The importance of weather-driven renewable energies for the United States energy portfolio is growing. The main perceived problems with weather-driven renewable energies are their intermittent nature, low power density, and high costs. In 2009 we began a large-scale investigation into the characteristics of weather-driven renewables. The project utilized the best available weather data assimilation model to compute high spatial and temporal resolution power datasets for the renewable resources of wind and solar PV. The weather model used is the Rapid Update Cycle for the years of 2006-2008. The team also collated a detailed electrical load dataset for the contiguous USA from the Federal Energy Regulatory Commission for the same three-year period. The coincident time series of electrical load and weather data allows the possibility of temporally correlated computations for optimal design over large geographic areas.
In the past two years, the team have designed and built a sophisticated mathematical optimization tool that is based upon linear programming with an economic objective. The tool has been constructed to include salient features of the electrical grid, such as; transmission, construction and siting constraints, reserve requirements, electrical losses due to transmission, asynchronous regions, "reliability" enforcement, capital costs, fuel costs, and many others.
One important test included existing US carbon free power sources, natural gas power when needed, and a High Voltage Direct Current power transmission network. This study shows that the costs and carbon emissions from an optimally designed national system decrease with geographic size. It shows that with achievable estimates of wind and solar generation costs, that the US could decrease its carbon emissions by up to 80% by the early 2030s, without an increase in electric costs. The key requirement would be a 48 state network of HVDC transmission, creating a national market for electricity not possible in the current AC grid.