TODAY’S STUDY: ECONOMIC VALUE OF VARIABLE GENERATION WITH INCREASING PENETRATION LEVELS
Strategies for Mitigating the Reduction in Economic Value of Variable Generation with Increasing Penetration Levels
Andrew Mills and Ryan Wiser March 2014 (Lawrence Berkeley National Laboratory)
Previously, Mills and Wiser  found a decline in the marginal economic value of different variable generation (VG) technologies with increasing penetration levels. Economic value in this case is primarily based on the avoided costs form other non-renewable power plants in the power system including capital investment cost, variable fuel, and variable operations and maintenance (O&M). That previous analysis, the “valuation report,” assumed only one VG technology was added at a time, that VG plants would be built at specific sites, that the commitment decisions for all thermal power plants except combustion turbines were fixed in the day-ahead market, that demand was largely inflexible, and that new bulk power storage facilities would be expensive to build. In this report, the “mitigation report,” we use the same model and data to evaluate individual options that have the potential to stem the decline in the marginal value of VG with increasing penetration levels.
We measure the effectiveness of mitigation measures by comparing the marginal value of VG once a mitigation measure is implemented to the marginal value in the Reference scenario based on the valuation report without the mitigation measure, and at the same VG penetration level. A positive change in the value of the VG technology with the mitigation measure indicates that implementing the measure increases the value above the value found in the valuation report for the same level of penetration.
Where the mitigation measures are found to increase the value of VG, we also want to understand if the mitigation measures themselves are more or less economically attractive with increasing penetration of VG. Ideally, a mitigation measure both increases the value of VG and becomes more economically attractive at the same time. We therefore develop metrics to assess the economic attractiveness of mitigation measures and examine the change in those metrics with increasing VG penetration. We only consider the benefits of mitigation measures, not the costs of implementing the measure, so our results only reflect part of the information required to conduct a full cost/benefit analysis of mitigation measures.
The valuation report examined four VG technologies: wind, single-axis tracking photovoltaics (PV), and concentrating solar power (CSP) with and without thermal energy storage. The changes in the value of PV and CSP without thermal storage were found to be largely similar and driven by the same factors. Adding thermal storage to a CSP facility was found to be an effective measure to mitigate the decline in the value of CSP with increasing penetrations. In this report, we therefore focus only on the effectiveness of mitigation measures for wind and PV. CSP with 6 hours of thermal storage is only included in the analysis as one of the potential mitigation measures to stem the decline in the value of wind or PV with increasing penetration.
The specific mitigation measures examined in the report include increased geographic diversity of wind siting, technological diversity (through simultaneous combinations of VG technologies), more- flexible new conventional generation, lower-cost bulk power storage, and price-elastic demand subject to real-time pricing (RTP).
To estimate the change in the marginal value of wind or PV with the mitigation measure, we implement the measure in the original model used in the valuation report. For each penetration level of wind or PV, we then develop new long-run investment decisions, new generation dispatch decisions, and new wholesale power market prices that reflect the impact of the mitigation measure. The wholesale power prices from this market in long-run equilibrium are then used to estimate the marginal value of additional wind or PV generation.
Only one mitigation measure is implemented at a time in this report. The change in the value of wind or PV if multiple mitigation measures were implemented simultaneously is not expected to be the same as the sum of the change in value with each mitigation measure implemented in isolation.
The economic attractiveness of the various mitigation measures with increasing wind and PV penetration uses the wholesale power prices from the Reference scenario. The specific metric used for each mitigation measure is described in more detail in the body of the report.
Additional relevant details of the different mitigation measures are as follows. The geographic diversity mitigation measure is based on siting wind plants in locations that minimize the variance of the aggregate wind production. This leads to wind sites being much more geographically distant from one another relative to the wind sites in the Reference scenario. The technological diversity scenarios involve adding 10% penetration of a different VG technology. In the case of PV, the value of PV with increasing penetration and 10% wind penetration is compared to the value of PV without any wind. We also examine the value of wind with 10% PV or 10% CSP with 6 hours of thermal storage. The more-flexible new conventional generation mitigation measure uses the assumption that new combined cycle gas turbines (CCGTs) can be started or stopped in real time as opposed to making all commitment decisions in the day-ahead market as was assumed in the Reference scenario. The low-cost bulk storage mitigation measure assumes that pumped-hydro storage with 10 hours of storage capacity can be built with a much lower investment cost than was assumed in the Reference scenario. Finally, the RTP mitigation measure assumes that load has a constant own-price elasticity of -0.1 and can change in the day-ahead and real-time market in response to changes in wholesale power prices.
Change in the Value of Wind with Mitigation Measures
The change in the marginal value of wind after implementing different mitigation measures is shown at different penetration levels in Table ES.1. The largest increase in the marginal value of wind at high penetration levels occurs with increased geographic diversity of wind sites (27% increase in the marginal value of wind relative to the value in the unmitigated scenario with 40% wind penetration), implementation of RTP (20% increase in the value of wind at 40% penetration), and availability of low-cost bulk power storage (11% increase in the value of wind at 40% penetration).
With 20% and 30% wind, the largest increase in the marginal value of wind is found with RTP. The increase in value is primarily due to an increase in the sum of the capacity and energy value of wind with increased penetration. The marginal value of wind increases because RTP tends to increase the load during times when wind power is available. Less than $2/MWh of the increase in the marginal value of wind with RTP is due to a decrease in the cost of day-ahead forecast errors.
The trends currently leading to the roll-out of smart meters and RTP programs are largely based on efforts to reduce peak demand, independent of mitigating changes in the economic value of wind. This analysis shows, however, that not only does wind increase in value with RTP, but also the attractiveness of RTP increases with increasing wind. A large portion of the increased attractiveness of RTP (as treated here) with increasing wind is derived from real-time response to events that were unforeseen in the day-ahead. Such active participation from the demand side through dynamic pricing programs as modeled here is a departure from the design of traditional demand-response programs and even some RTP programs as they are currently implemented.
With 40% wind, the largest increase in the value of wind is with increased geographic diversity. Whereas the selection of wind sites in the Reference scenario is based on a number of factors, in the highgeographic-diversity scenario the selection of wind sites is based entirely on minimizing the total variability of the aggregate wind production. Picking these sites increases the marginal value of wind at 40% penetration by $10.6/MWh.
The increase in the marginal value of wind at 40% penetration with increased geographic diversity is based on an increase in the energy value of wind, a smaller increase in the capacity value, and a small reduction in the cost of day-ahead forecast errors. Wind from diverse sites is less correlated with wind from more concentrated sites. Increasing geographic diversity, therefore, reduces the frequency of extremes: diverse wind tends to generate at different times rather than all sites generating at the same time or no wind generating at a given time.
Though not shown in Table ES.1, we also examined a case with very low geographic diversity in wind siting. When tightly clustered wind is generating strongly from all sites at once, conventional generation with lower and lower variable costs begins to be displaced. This lowers wholesale prices during times with significant wind. Wind from geographically diverse sites can earn higher revenue compared to tightly clustered wind. Concentrating wind in one region decreases the marginal value of wind by $6/MWh. The decrease in value compared to the original siting is driven primarily by an increase in the cost of day-ahead forecast errors.
The increased attractiveness of wind at diverse sites must be compared to factors that may increase the costs of wind sited in this fashion, including the potential for higher transmission costs and lower-quality wind resources. Since the cost of wind can vary greatly depending on the local wind resource quality, geographic diversity is not likely to dominate siting decisions.
Finally, while the technological diversity scenarios—10% PV and 10% CSP6—do not greatly increase the marginal value of wind at 20% penetration or higher, the results are important in that adding these solar technologies does not substantially decrease the value of wind at 20% and 30% penetration. These results suggest that, if a full comparison of costs and benefits could justify 10% PV penetration alone or 20% wind alone, then 30% penetration from a combination of wind and PV could be similarly justified. The remaining results regarding the impact of low-cost storage and more flexible CCGT’s are discussed in the main report. The negligible effect of more flexible new CCGT’s is due in part to the flexibility already available from existing resources. Flexibility of new CCGTs would play a larger role in areas that lack as much flexibility from existing resources.
Change in the Value of PV with Mitigation Measures
The change in the marginal value of PV after implementing several different mitigation measures is not the same as for wind, Table ES.2. By far, the largest increase in the marginal value of PV at high penetration levels occurs with the availability of low-cost bulk power storage. With 30% PV penetration, the availability of low-cost storage increases the value of PV by nearly $20/MWh (an 80% increase in the marginal value of PV relative to the value in the unmitigated scenario with 30% PV penetration), due primarily to an increase in the energy value of PV. With high PV penetration, storage is charging during times with PV generation, which increases prices during these times relative to their level absent any new storage.
Likewise, the marginal value of bulk power storage increases with high PV penetration. With 30% penetration of PV, the marginal value of new storage increases by over $100/kW-yr relative to the value of storage with no PV. Decreases in the cost of multiple-hour bulk power storage could make storage the most attractive mitigation measure for moderating the decline in the value of PV at very high penetration levels. For more modest penetration levels of PV, the two most effective mitigation measures are RTP and technological diversity with 10% wind penetration. At 10% PV penetration, these two measures increase the value of PV by more than double the increase in the value of PV from low-cost storage.
The increase in the marginal value of PV with 10% wind penetration is due to an increase in the capacity value of PV relative to the Reference scenario. The value of wind also increases with 10% PV penetration, suggesting that if wind were attractive without PV it would be just as attractive, if not more, with 10% PV. With higher PV penetration, however, 10% wind begins to decrease the value of PV relative to the Reference scenario. Wind is therefore an attractive mitigation measure for relatively modest PV penetration levels. CSP6 decreases the value of PV, so it is not considered in detail as a mitigation measure.
The increase in the marginal value of PV with RTP is due to an increase in demand when PV is generating. As mentioned earlier for wind, the development of RTP programs is currently largely independent of issues involving renewable energy, but the attractiveness of RTP does increase with both increasing PV and wind penetration. The remaining results regarding the impact of more flexible CCGT’s are discussed in the main report.
In summary, several mitigation measures both increase in attractiveness with increasing penetration of wind and PV and increase the marginal value of wind and PV relative to a scenario without the mitigation measure. This report is also helpful in highlighting measures that may not increase the value of wind or PV or, in some cases, can even decrease the value. The largest increase in the value of wind comes from increased geographic diversity. The largest increase in the value of PV comes from assuming that low-cost bulk power storage is an investment option. Other attractive options include RTP and technology diversity. These mitigation measures may have costs or may be driven by factors not directly related to increases in wind and PV. Decisions to implement specific mitigation measures should account for these costs and consider other important factors not included in this analysis