While there are many options to increase grid flexibility, in this work we focus on the potential use of CSP with TES. Thermal storage extends the contribution of solar electricity generation by shifting generation to im­prove its coincidence with normal demand, and by improving system flex­ibility. The latter is accomplished by reducing constraints of ramping and minimum generation levels.

CSP was added to REFlex using hourly generation values produced by SAM. SAM uses the direct normal irradiance (DNI) to calculate the hourly electrical output of a wet-cooled trough plant (Wagner and Gilman 2011). The choice of technology was based primarily on data availability at the time of analysis as opposed to any presumption regarding CSP technology or economics. The results should be applicable to any CSP technology able to deploy multiple hours of thermal energy storage. For our base case, we assume 8 hours of storage and that the electrical energy produced by the plant can be dispatched with an effective 95% efficiency. In this initial analysis we did not consider the effects of part loading or multiple starts on plant efficiency. Distribution of locations was based on the study de­scribed by Brinkman et al. (2011).

Figure 7 illustrates the importance of dispatchability at high solar pen­etration. This scenario is identical to Figure 4, except PV provides 15% of annual demand and CSP meets 10% (so the contribution of solar tech­nologies in total is greater in the PV/CSP case in Figure 7). The figure shows two CSP profiles. This first “non-dispatched CSP” is the output of CSP if it did not have thermal storage. It aligns with PV production, and would result in significant solar curtailment. The other curve is the actual dispatched CSP, showing its response to the net demand pattern after wind and PV generation is considered. It shows how a large fraction of the CSP energy is shifted toward the end of the day. In the first day, this ability to shift energy eliminates curtailment. On the other days, the wind and PV

resources exceed the “usable” demand for energy in the early part of the day, resulting in curtailed energy even while the CSP plant is storing 100% of thermal energy. However, overall curtailment is greatly reduced. Solar technologies provide an additional 5% of the system’s annual energy com­pared to the case in Figure 4, but the actual annual curtailment has been reduced to less than 2%, including the losses in thermal storage.

Figure 8 shows how the addition of CSP/TES can increase the overall penetration of solar by moving energy from periods of low net demand in the middle of the day to morning or evening. In this figure there is an equal mix of CSP and PV on an energy basis and the PV-only curves are identi­cal to those in Figure 5.

Figure 8 demonstrates the importance of dispatchability to reduce cur­tailment and increase the overall penetration of solar via the ability to shift solar energy over time. However, the analysis to this point assumes that CSP and PV are complementary only in their ability to serve different parts of the demand pattern. We have not yet considered the additional benefits of CSP to provide system flexibility by replacing baseload genera­tors and generators online to provide operating reserves.

The importance of system flexibility can be observed in Figure 4, where conventional generators must ramp up rapidly to address the decreased output of PV during peak demand periods. In order to meet this ramp rate and range (along with sufficient operating reserves) a significant number of thermal generators will likely need to be operating a part-load, creating a minimum generation constraint during periods of solar high output. This is represented by the flat line occurring in the middle of each day when the aggregated generator fleet is at their minimum generation point. Compar­ing the CSP/PV case in Figure 7 to the PV only case in Figure 4, we see that the CSP is dispatched to meet the peak demand in the late afternoon/ early evening, and the overall ramp rate and range is substantially reduced. In Figure 4 conventional generators need to ramp from about 18 GW to over 45 GW in just a few hours, while in Figure 7 the generators need to ramp from 18 GW to less than 30 GW.

Adding a highly flexible generator such as CSP/TES can potentially reduce the minimum generation constraint in the system. In the near term, this means that fewer conventional generators will be needed to operate at part load during periods of high solar output. In the longer term, the ability


FIGURE 7: Simulated system dispatch on April 7-10 with 15% contribution from PV and 10% from dispatchable CSP


Подпись: Solar Curatilment Rate

0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50%

Fraction of Energy From Solar



of CSP to provide firm system capacity could replace retiring inflexible baseload generators.

CSP plants with TES add system flexibility because of their large ramp rate and range relative to large baseload generators. Many CSP plants, both existing and proposed, are essentially small steam (Rankine-cycle) plants whose “fuel” is concentrated solar energy. Few of these plants are deployed, so it is not possible to determine their performance with abso­lute certainty. However, historical performance of the SEGS VI power plant provides some indication of CSP flexibility. Figure 9 provides a heat rate curve based on an hourly simulation model to assess the performance of parabolic trough systems, and validated by comparing the modeled out­put results with actual plant operating data (Price 2003). It indicates a typical operating range over 75% of capacity, with only a 5% increase in heat rate at 50% load. Figure 9 also provides historical data from small gas-fired steam plants which also indicates high ramp rate and range and fairly small decrease in efficiency at part load (about a 6% increase in heat


rate at 50% load). These plants also often operate as low as 25% of capac­ity, although with lower efficiency. This provides a strong indication that CSP plants should be able to provide high flexibility.

The change in minimum generation constraints is dependent on both the flexibility of CSP plants and the flexibility of generators supple­mented or replaced with CSP. As discussed previously, nuclear plants are rarely cycled in the United States, while coal plants are typically oper­ated in the range of 50%-100%. Because it is not possible to determine the exact mix of generators that would be replaced in high renewables scenarios, we consider a range of possible changes in the minimum generation constraints resulting from CSP deployment. For example, deployment of a CSP plant which can operate over 75% of its capaci­ty range could allow the de-commitment of a coal plant which normal operates over 50% of its range. In this scenario each unit of CSP could reduce the minimum generation constraint by 25% of the plant’s capac­ity. This very simplistic assumption illustrates how the dispatchability of a CSP plant should allow for a lower minimum generation constraint.



FIGURE 11: Curtailment of solar assuming an equal mix (on an energy basis) of PV and CSP and impact of CSP grid flexibility

Reducing this constraint should allow for greater use of wind and PV. As a result, as CSP is added, the system can actually accommodate more PV than in a system without CSP.

This is illustrated conceptually in Figure 10, which shows the same 4-day period as in Figures 4 and 7. CSP still provides 10% of the system’s annual energy, but now we assume that the use of CSP allows for a de­creased minimum generation point, and the decrease is equal to 25% of the installed CSP capacity. In this case about 21 GW of CSP reduces the minimum generation point from about 18 GW to 13 GW. This generation “headroom” allows for greater use of PV, and enough PV has been added to meet 25% of demand (up from 15% in Figure 7). As a result, the total solar contribution is now 35% of demand, significantly greater than the PV-only case shown in Figure 4, and total curtailment is less than the 6% rate seen in Figure 4. By shifting energy over time and increasing grid


FIGURE 12: Increase in PV penetration as a function of CSP penetration assuming a maximum PV marginal curtailment rate of 20%. CSP flexibility is defined as the fraction of the CSP rated capacity that is assumed to reduce the system minimum generation constraint.

flexibility, CSP enables greater overall solar penetration AND greater pen­etration of PV.

Figures 11 and 12 show the potential overall impact of the flexibility introduced by CSP and the corresponding opportunities for increased use of PV. Figure 11 builds on Figure 8 by adding the flexibility benefits of CSP. The figure assumes that each unit of CSP reduces the minimum gen­eration constraint by 25% of its capacity, and an equal mix of PV and CSP on an energy basis. In this case, the addition of CSP allows PV to provide 25% of the system’s energy with very low levels of curtailment.

Figure 12 more directly illustrates the relationship between the reduc­tion in minimum generation constraint and potential increase in PV pen­etration. The figure shows how much more PV could be incorporated at a constant marginal curtailment rate of 20% when CSP is added. In this scenario, the x-axis represents the fraction of annual system energy pro­vided by CSP. Increased penetration of CSP results in a linear decrease in minimum generation constraints. The figure illustrates two CSP flexibility cases. In one, each unit of CSP reduces the minimum generation constraint by 20% of its capacity; in the other, the rate of reduction is 40%. These amounts are not meant to be definitive, but represent a possible impact of CSP in reducing minimum generation constraints.

Overall, this analysis suggests that CSP can significantly increase grid flexibility by providing firm system capacity with a high ramp rate and range and acceptable part-load operation. Greater grid flexibility could in­crease the contribution of renewable resources like solar and wind. This demonstrates that CSP can actually be complementary to PV, not only by adding solar generation during periods of low sun, but by actually en­abling more PV generation during the day. This analysis also suggests a pathway to more definitively assess the ability of CSP to act as an “en­abling” technology for wind and solar generation.