A systematic evaluation of wind's capacity credit in the Western United States

Abstract

The degree to which wind energy can contribute to the capacity needed to meet resource adequacy requirements, also known as capacity credit (CC), can be impacted by many factors. The CC varies regionally with wind resource and correlation to net load and may also vary with wind technology (land‐based versus offshore) and turbine specifications. We use a probabilistic resource adequacy tool to systematically assess the CC of multiple technologies for near‐term wind deployment across the Western US power system to help inform planning decisions. We find that wind CC varies by weather year and location but averages 16% for land‐based turbines and 41% for offshore turbines, with large regional variations. The average wind CC increases to 20% for land‐based and 53% for offshore when considering only the sites in the top 25th percentile of capacity factor. The CC of land‐based wind is generally lower than its capacity factor, which indicates a low correlation of generation to system needs. However, offshore wind shows a substantially higher CC, which is driven by not only a higher capacity factor but also a greater correlation with system needs.

Comparing Capacity Credit Calculations for Wind: A Case Study in Texas

“…To conduct a probabilistic resource adequacy analysis, we apply NREL's Probabilistic Resource Adequacy Suite (PRAS) tool (Stephen 2019). PRAS uses interregional power transfer constraints and Monte Carlo draws from probability distributions for generator availability to calculate the loss of load resulting from supply shortfall or transmission constraints 15 (Beiter et al 2020, Frew et al 2019, Stephen, Hale, and Cowiestoll 2020, Stephen 2021. For each hour, we use 100,000 Monte Carlo draws of forced generator outages to compute the EUE for a given level of net load 16 , which are aggregated into annual values.…”

“…The hourly profiles for existing wind, utility-scale PV, and distributed PV are from the Wind Integration National Dataset (WIND) Toolkit (Draxl et al 2015) and the National Solar Radiation Database (Sengupta et al 2018), processed through NREL's System Advisor Model (Blair et al 2018) with specific technology assumptions to generate CF profiles (Maclaurin et al 2019). For representing new wind resource, we also use the WIND Toolkit, which contains wind data for 2-km by 2-km grid cells for this footprint over 7 weather years (Draxl et al 2015); however, to reduce computational demands and to more closely match the footprint of modern wind power plants, we aggregated profiles to a resolution of 20 km by 20 km 9 (Jorgenson et al 2021). We consider two turbine types: one technology for land-based sites and one for offshore sites.…”

“…We consider two turbine types: one technology for land-based sites and one for offshore sites. Table 2 shows the technology specifications we use to generate the profiles for the incremental wind (Jorgenson et al 2021) 10 . We employ 7 years of system load, wind, and solar profiles (2007 through 2013).…”

“…Two main methods are used to estimate CC 1 : reliability-based and approximation-based methods (Sioshansi, Madaeni, and Denholm 2013). Reliability-based methods typically use resource adequacy models to calculate metrics such as the loss-of-load probability (LOLP) or expected unserved energy (EUE) given inputs such as hourly loads, generator information, and often transmission network data (Corporation 2018, Jorgenson et al 2021, Mills and Rodriguez 2020. These methods can calculate the CC of an incremental resource by computing the reliability metric before and after the addition of the new resource, with any change in the metric attributed to the incremental resource.…”

“…As an example, a stand-alone PV plant could have a CC of 20%; however, if capacity was needed during nighttime, it would not be available at all. In addition, CC is only one factor that might influence the deployment of various generator types (including wind), and it can often be lower priority than selecting high CF resource areas or transmission access, which can be important drivers to capital cost (Jorgenson et al 2021). In short, calculating CC is a useful but imperfect means to incorporate weather-dependent resources into established resource adequacy assessment and integrated planning processes.…”