Impact of Time Resolution on the Projected Rates of System Penetration by Intermittent Generation Technologies

To hedge against the limited resources of fossil fuels and to reduce the emissions of green house gases, it is expected that our future electricity system will include more intermittent technologies, including wind and PV. To better understand how to develop energy systems that rely on intermittents, systems models are used to assess the cost at which intermittents become competitive, the degree of penetration as their costs are reduced, their impact on the optimal structure of the balance of the system, and their affect on total system costs. Modeling approaches designed for dispatchable technologies are not entirely appropriate for modeling intermittent technologies, since they, naturally, assume that generation can always be dispatched to meet demand. Intermittent generation cannot be dispatched--its output varies from hour to hour and from day to day on its own schedule, heedless to system needs. This research assesses the difference in results associated with the different approaches to modeling intermittency. The analyses compare cases using the hourly loads and intermittent generation patterns, cases in which the loads and generation were averaged over several hours, and cases in which the loads and/or the generation were represented by the annual averaging scheme used in the National Energy Modeling System developed by the Energy Information Administration. Three significant characteristics of an intermittent generator are the average power production (capacity factor), the coincidence of its power production and loads, and the variation in the magnitude of its power production. Economic models of the energy system represent these characteristics with differing degrees of accuracy. It is expected that different representations of the characteristics of an intermittent generator will give different answers to the sorts of questions posed above. This research assesses the magnitude and types of errors that are introduced by not representing the characteristics of the intermittents accurately. The most accurate representation of an intermittent generator uses its actual output from moment to moment. Here we use a one hour resolution over a full year of generation as the base case. This captures the variations from hour to hour and day-to-day. However, some energy modeling systems are based on a load duration curve approach for characterizing the variation in energy demand. This is quite suitable for dispatchable technologies since the generators can always be dispatched to meet the load whenever it occurs. When an intermittent generator is represented in this structure, it is represented as having a constant output equal to its capacity factor over long intervals (many hours). This approach captures the capacity factor of the intermittent and to some extent it can capture the coincidence of generation and demand, but does not capture the effect of the short term variations in output. In this paper, we evaluate the impacts of time resolution on the economic evaluation of wind and solar PV within a simple energy system. We assess the penetration of each intermittent generator as its cost is decreased. At the same time, the model optimally readjusts the capacities and dispatch of the conventional generators as the intermittent technology penetrates. This investigation compares the trajectories of intermittent penetration under a several different representations of intermittent generation and demand. In the following sections, we first discuss the approach to analysis, for both the load duration curve approach to representing intermittent generation and several averaging schemes. We then present results and conclusions