The future of the utility industry has become a central focus for many as the sector grapples with several existential threats. Among the chief threats looming on the horizon is the large projected growth in distributed energy resources (DERs) and its potential to compound the impacts of the anemic growth in net load observed in many regions today. But this growth in DERs is relatively recent.
While the resource base has certainly grown significantly for specific resources in particular regions such as the solar photovoltaic generation in California or the demand response in PJM, on a national basis, these resources still occupy a relatively small fraction of the overall mix. Nevertheless, the conditions for growth for this class of resource are approaching a tipping point toward widespread viability in many more markets, and there is growing enthusiasm around the potential for growth of DERs in the years and decades to come.
The growth of distributed generation (DG) and its impact on price formation in U.S. capacity markets implies a fundamental shift in the structure of resource adequacy mechanisms. As variable, distributed generation increasingly becomes a prevalent source of generation in regions, changes in capacity market dynamics will have a profound impact on generating assets and their future economic viability. Without commensurate changes in capacity market structure to account for these changes, system reliability will be compromised.
Despite a great deal of discussion around DER trends and their prospects for growth, the impact of these resources on generation assets and price formation in the market has been largely overlooked. The impacts of these resources on system operation and load growth have received a great deal of attention through the volumes of integration and impact studies in literature. But while the impact on the diurnal load profile and the continuing uproar surrounding net metering policies have received extensive media interest, the sometimes subtle discussion of the impact of these resources on market structure and on the future viability of generating assets more broadly has not received the attention it deserves.
Because of the geographic distribution of distributed PV resources, their place in the power system topology and the variable nature of their output, the traditional market constructs built around fixed-load and dispatchable power begin to break down in fundamental ways. While some might characterize this as the triumph of new technologies over the power plants of the past, the truth is much more complex, and the stakes for addressing these changes to the market are much higher than such a simple narrative suggests. In fact, the ability of the markets to ensure reliable power delivery rests on the efficient operation of these markets, and so the need to find solutions to these issues is far more important than the relative economics of any given fuel source or technology.
These market impacts will be felt most acutely in organized markets with well-developed capacity market mechanisms. In those regions, the adequacy of system resources for meeting the demand for electricity depends on effective mechanisms for price formation to provide market signals to incent investment in new resources when conditions of supply and demand warrant it. These mechanisms are critical to both the integrity of the market and to the maintenance of resource adequacy. To understand the impact of distributed energy resources on price formation, it is important to understand not only the scale of these resources but their output characteristics and their place in the power system topology.
Sources of generation like wind and solar with a free fuel source will operate at zero marginal cost – operations and maintenance costs are effectively zero. This leads to a price suppression effect, wherein zero-fuel-cost plants flatten the left side of the supply curve, resulting in lower wholesale energy prices in the real-time market. Real-time prices can descend into negative territory – as happened in Texas – when a production-based incentive drives the effective short-run marginal cost below zero and there is insufficient transmission capacity to move power to load pockets.
These impacts alone have the impact of distorting capacity prices without mitigating measures, and, in fact, we have seen that the reduction of infra-marginal revenues in Europe have led to the recent shutdown or mothballing of 30 GW of gas-fired capacity, including last year's decision by E.On SE to mothball a two-year-old combined cycle unit in Malzenice, Slovakia.
Unlike utility-scale systems that can rely on the bulk power grid to more effectively leverage geographic diversity of the resource, DG resources interconnect at the distribution level where the impacts of variability are not as easily mitigated.
The aforementioned negative prices in Texas were effectively alleviated in large part through large-scale transmission expansion and the development of the competitive renewable energy zone (CREZ) transmission lines. In contrast to this, DERs exist at the low-voltage side of the power system and, therefore, do not have the same level of access to the bulk grid and its ability to transfer power across great distances. Distribution feeders function as the capillaries of the power system, and with the uneven geographic distribution of DERs leading to heavy concentration of systems on individual feeders, the ability to leverage geographic diversity of the resource across weather fronts and climatic zones becomes a much greater challenge. This further exacerbates the price suppression phenomenon described above and creates the need for other balancing measures and quickly responding resources.
The critical question from a market integrity standpoint is how do the price suppression impacts and the variability of distributed resources impact price formation in the markets in general and in the capacity markets in particular? For those markets such as New England and PJM that rely on three-year capacity market constructs to maintain resource adequacy, market fundamentals rely on the assumptions of an accurate valuation of resource contribution to loss of load expectation and an accurate reflection of supply-and-demand dynamics.
Distributed PV and PJM capacity prices
The PJM market operates one of the most well-developed capacity market constructs in the U.S. and provides a unique window into how DERs could substantially call into question both the assumptions of accurate resource valuation and price signals accurately reflecting the market's supply/demand balance.
PJM's reliability pricing model (RPM) – its three-year capacity market – relies on the mechanism of net cost of new entry (CONE) to form the upper boundary of the market demand curve – the so-called variable resource requirement curve – that dictates the clearing price in the base residual auction (BRA). Because net CONE is defined as the cost of adding a new resource minus the expected energy revenue from that generator, the impact of cost suppression will be to inflate net CONE.
In fact, the ability of distributed generators to underbid all conventional generation in economic dispatch on the basis of near-zero, short-run marginal cost means that expected capacity factors for all plants will be lower in those regions where DER penetrations are high. The subsequent impact on bidding behavior across the market would be to inflate capacity prices to unsustainable levels.
The recently released PJM renewable integration study (PRIS) looked at 10 scenarios for the market through 2026 at varying levels of wind and solar deployment with levels of distributed solar energy resources exceeding 30 GW through 2026 in the PJM footprint. That level of deployment represents roughly 18% of the peak demand in the PJM footprint in that year. In places like New Jersey and Maryland with favorable policies and rate environments for PV and where the market for PV has outpaced the rest of the PJM region, one could expect substantially higher capacity penetrations locally under such scenarios. Furthermore, the saturation in specific PJM locational marginal pricing zones and specific feeders could even exceed total peak load several times over to the extent that solutions are in place to accommodate high-volume bidirectional power flow at the distribution level. This means that without proper planning, price supports and other incentives could drive local penetration levels to unprecedented levels.
Under these types of conditions, it is not unreasonable to expect significant impacts on energy prices in the real-time market with significant commensurate impacts on the clearing price in RPM. This impact is especially high for distributed resources because, unlike utility-scale resources that can bid into the capacity market on the basis of their impact on loss of load expectation (i.e., their effective load carrying capability [ELCC]), distributed resources do not participate in capacity procurement for the regional transmission organization. Therefore, while systems interconnecting to the bulk system can bid in an amount of unforced capacity equal to the determined capacity value – central PV, for example, was found to have an ELCC of 62%-66% – the distributed resources cannot bid their capacity in the BRA, which means that they are effectively inflating the demand curve through price suppression without offsetting that by extending the capacity supply curve.
While price distortions such as these are counterproductive for effective market operation, there might be a tendency to view high-capacity prices as a positive price signal incenting new entry into the market, especially in the current capacity-long environment and in light of the most recent BRA for the 2016-2017 delivery year, which saw prices collapsing in specific regions.
This instinct should be tempered for two reasons: Price formation that diverges from fundamental supply/demand balance will produce inefficient allocation of capital in the marketplace. This trend is also coupled with an inability of the current capacity market constructs to fully capture the value of all resource attributes, which means that high capacity prices could produce not only the wrong amount of capacity but the wrong type of resources as well.
Resource valuation and flexibility concerns are central to price formation in the market. It is clear that in addition to the capacity market distortions already present due to the price suppression, if the need for flexibility and fast ramping resources is not somehow internalized into the capacity procurement mechanism, then the resource adequacy objective function is fundamentally incomplete. That means that in addition to inefficient capital allocation and resource mismatch, the system operator's view of resource adequacy is fundamentally incomplete because the traditional assumptions around dispatch and the relationship between nameplate capacity and loss of load expectation are undermined as the participation of variable energy resources, and distributed generation in particular, continues to increase.
This article was excerpted with permission from the white paper ‘Distributed Generation's Future Impact on the U.S. Capacity Markets.’ The full report, which includes a discussion about how California's experience in variable generation is and is not applicable to the eastern U.S., can be found here.
Samir Succar is a technical specialist with ICF International, a market research firm and consultancy based in Fairfax, Va. He analyzes and models power market supply-demand fundamentals, develops power market price forecasts, and performs generation asset valuations. Reach him by email at email@example.com.