Estimating the hourly (or subhourly) state of power plants with various fuel types and costs for plausible wind penetration scenarios is both important and difficult. Most of the recent wind integration studies use production cost simulation (PCS) software for this purpose. Most of the studies used proprietary PCS tools such as GE MAPS (Charles River Associates, 2010; GE Energy, 2005a, 2008, 2010), Ventyx PROMOD (EnerNex Corporation, 2006, 2011), or Global Energy’s PROSYM (EWEA, 2009). The most sophisticated of these model generator ramp rates, startup/ shutdown costs, and transmission limits (typically using the DC power flow model). One study (EWEA, 2009) used a research PCS tool, WILMAR (Weber, Meibom, Barth, & Brand, 2009), which was specifically designed for wind integration studies. Each of these tools uses a unit-commitment generator cost model. The most sophisticated models (e. g., PROMOD) solve for power flows and account for transmission constraints.
The evolution of production cost simulations is evident in these studies. The NREL 2030 study (U. S. DOE, 2008) used an internally developed PCS model, with detailed cost data for a large area (the whole of the eastern United States) but (as noted previously) a transportation model of electrical flows. The newer studies include substantial detail regarding generator and transmission constraints. Since most of the newer studies used PCS models for 1 to 3 years of data, with 1- to 3-hour time increments, the results provide reasonable estimates of the effect of wind power on dispatch and unit commitment costs. Because the time scales for unit commitment calculations are somewhat longer, the reduced higher-frequency variance found in mesoscale data (Sec. 3.2) should have little effect on unit commitment calculations.
One area in which PCS technology is rapidly evolving (for both PCS and the industrial unit commitment systems that PCS simulates) is the ability to deal with the stochastic nature of renewable generation. The science of stochastic unit commitment is rapidly maturing (see, e. g., Takriti, Birge, & Long, 1996; Wu et al., 2007), and there is increasing evidence that effective use of this technology can reduce wind integration costs (Madaeni & Sioshansi, 2013). As stochastic methods are effectively adopted in the PCS software used in integration studies, future studies are likely to provide more detailed insight into the benefits, costs, and challenges of wind integration.