A version of this article was published on Greentech Media on March 23, 2015.
By Eric Gimon, Robbie Orvis, Sonia Aggarwal and the experts of America’s Power Plan
The electric grid is undergoing a rapid metamorphosis as renewables become a significant part of the system. The last five years have seen fast growth of both solar and wind. In fact, more than six gigawatts of solar capacity were installed in 2014, 30 percent more than the prior year’s solar installations and sixteen times the capacity installed five years before. Wind represented just over four percent of total US power generation, and an even bigger fraction in large portions of the country. These new technologies are entering a system designed around a very different set of resources. As low-cost renewables provide a growing share of the electricity, grid operations—and thus power markets, financial structures, and policies—must evolve. Indeed, some already have; this article draws lessons from power contracts and markets about the evolving role of renewables from two different organized markets.
Wind power in a restructured market
In restructured markets, the resource with the lowest marginal cost should be dispatched first (including demand-side resources like demand response or efficiency). The resource mix is also “security constrained,” meaning that the system operators have to make sure that the portfolio of resources dispatched will work reliably within existing transmission system constraints. If many resources with the same marginal cost, for example multiple wind farms, find themselves behind a transmission bottleneck, security constraints will mean they cannot all be dispatched at their maximum capacity and some will need to be turned down or “curtailed,” either voluntarily or forcefully. Reducing their generation in this way means they’re paid less even though they’re available to the system, so market operators must look for the most efficient way to minimize these economic losses while maintaining system reliability.
Today, most renewable energy projects are financed and contracted using Power Purchase Agreements (PPAs). Under most PPAs, project developers commit to design, permit, finance, and operate a renewable energy power plant in exchange for a commitment by the customer, often referred to as an “offtaker,” to buy all the energy produced at a fixed rate over a long time horizon, typically ten to 25 years. Because PPAs are structured around the amount of electricity generated, anything that affects the total electricity output, like curtailment, represents a significant risk for the developer. Long-term contracts for variable resources rely on careful analysis of the wind and solar resource and probabilities about the average number of hours they’ll be available in a given year, and increasingly the amount of expected curtailment. Anything that reduces the risk by improving the predictability of wind and solar availability—and the number of hours they expect to get paid for their generation—is a plus for all parties to the contract.
The Midwest approach to increasing predictability: Economic dispatch for renewable power
In a move to improve its operating environment, dispatch more efficiently, and provide a more stable context for PPA-driven wind resources to generate revenue, the Midwest Independent System Operator (MISO) switched to “economic” dispatch for wind plants in 2011. MISO’s Dispatchable Intermittent Resource (DIR) tariff requires wind plants that began operating in April 2005 or later to offer their energy into the real-time market and participate in security-constrained economic dispatch. Before the DIR protocol, wind plants were manually curtailed when needed (in practice, grid operators picked up the phone to call wind operators and ask them to turn down), resulting in an inefficient system and relatively high curtailment rates – e.g., 3.7 percent of total wind generation. Wind plants also tended to incur significant penalties for under-delivery when real-time conditions did not match their day-ahead forecast.
The shift to economic dispatch means that wind plants can now be turned down according to the market, rather than manually curtailed via phone calls from MISO. The transition has significantly reduced manual curtailment, with the most recent data showing only 0.2 percent of total generation curtailed in this way. However, it’s very important to note that “down dispatched” wind electricity – this new market-based curtailment – is still significant, and overall curtailments of wind resources in MISO have only been modestly reduced since DIR was adopted. On the other hand, given that manual curtailment was going to happen anyway, economic dispatch can actually help reduce uncertainty in the financial returns from variable generation. On top of that, wind plants that are registered as DIRs can submit day-ahead offers that get cleared in the day ahead market and integrated into the MISO dispatch, and the wind plants can better manage penalties when real-time performance does not match day-ahead predictions because they can update their forecasts on short intervals in real time as they approach the dispatch window. Meanwhile, transitioning renewables to economic dispatch in MISO has helped address flexibility and reliability concerns and gives grid operators more control.
In MISO and other organized markets, PPAs are mostly “synthetic.” This means that they do not govern a physical shipment of electrons from the producer to the offtaker. In reality, energy is sold at the nearest regional pricing point (“node”) to the producer, bought perhaps at the same node or at the one nearest to the offtaker’s customers, and then a “contract-for-differences” ensures that each party is made whole. The terms of the contract may also address how any price differences experienced throughout this process are handled given the agreed-upon contract price. Sometimes the offtaker is not even the final user of the electricity, but simply an intermediate power marketer that sells the electricity on to another customer. Given the financial—not physical—nature of these PPAs, it makes a great deal of sense to integrate wind plant dispatch more coherently into the organized market. The combination of plentiful wind resources, tax advantages, and a friendly environmental policy environment has allowed wind developers to thrive in the large MISO footprint, and wind now makes up about seven percent of MISO’s generation.
In the near term, the MISO DIR makes the financials for variable generators more predictable—driving down capital costs for the projects—and makes use of the advantages of economic dispatch to efficiently manage system reliability. In the long term, though, that means the MISO states must come to terms with renewable curtailment, foregoing nearly-free, zero-carbon electricity production from already-built projects. Luckily, there are solutions available today to drive down curtailment: more physical transmission between balancing areas or constrained locations within a balancing area, increased financial trading between balancing areas, or other flexibility resources like price-responsive demand.
Even without transmission constraints, though, moving all generators to economic dispatch raises other longer-term questions. For example, how would economic dispatch work when the market as a whole is flooded with zero marginal cost resources? More specifically, how would grid operators determine which plants get dispatched and which do not when the marginal generating unit repeatedly bids almost the same price as many other units over long periods of time? While not an issue today with relatively low levels of curtailment, these issues will become important in the future.
The Western approach to increasing predictability: New kinds of power contracts
While including wind generators directly in economic dispatch seems like a good solution in the MISO system, things are somewhat different out west. In California, developers have become wary of the potential increase in renewable curtailment and are starting to structure new contracts that limit curtailment to a fixed amount (e.g., five percent of potential output, still greater than the 3.7 percent that has been seen in MISO). With these new contracts in place, grid operators can curtail beyond this amount, but they are required to compensate developers as if they had not. As it turns out, a variety of differences between MISO and California underlie these different approaches; understanding these dynamics provides lessons for grid operators and policymakers in other parts of the country.
First, the makeup of California’s electricity system is very different from MISO’s grid. The two regions have similar levels of wind energy penetration (seven percent), but California also integrates significant amounts of in-state hydro resources (10-20 percent) along with solar (five percent), biomass (three percent), and geothermal (six percent), which enable the state to meet its 2020 goal of 33 percent renewable energy. Additionally, within the California Independent System Operator (CAISO) footprint, renewable procurement is dominated by just three principal offtakers, the three big investor-owned utilities. PPAs typically turn over effective control of the generation output to this limited number of utility offtakers—essentially monopsonies. The utilities optimize the portfolio of resources they own or contract with under PPAs, and all those resources are under the control of a scheduling coordinator who is the main point of contact with the CAISO. They curtail variable generation as a function of their needs. For example, a utility may have already pre-purchased fuel for delivery on a given day, and may choose to burn that fuel in one of the natural gas plants it controls, rather than schedule the full output from a variable generator. Under these circumstances, it is easy to see how a utility’s incentive to optimize its portfolio of resources may not be aligned with a renewable project developer’s incentive to maximize revenue, thus creating a degree of risk to be addressed in contract structures.
Second, California’s grid is comparatively inflexible during some periods of peak renewable generation. This lack of flexibility comes from many places, including: legacy contracts that protect some resources that could otherwise be turned down, local capacity requirements, the way with which imports are dealt in the market, and barriers to balancing resources inside the CAISO territory with adjacent systems (e.g. LADWP). Comparing two recent grid integration studies—one conducted for California’s investor owned utilities and one conducted for a group of advanced energy companies—provides an apt illustration of the flexibility challenge. By varying some key assumptions on local capacity requirements, resource mix, and interactions with other states, the utilities showed a 50 percent non-hydro renewables scenario with 4-9 percent curtailment in 2030 while the advanced energy companies showed a 55 percent non-hydro renewables scenario with only 1-1.5 percent curtailment.
Given the context above, the California Independent System Operator has suggested that significant curtailment might be necessary in certain periods by the year 2020. Unlike in MISO, where real-time market prices are effectively reducing output from wind as needed, California is unsure that operators will curtail similarly. Given this problem, CAISO has proposed dropping the bid floor for economic dispatch in the real-time market to negative $300/MWh (by contrast, MISO gets the down-dispatch they need with prices typically in the negative $11/MWh range). The use of such a low floor is intended to drive operators to curtail when it is needed or else face significant penalties. Significant excursions into this kind of price territory pose significant risks to developers of new renewable energy resources, hence the contractual protections.
Lessons for policy makers
What lessons can policy makers take from the differences between the MISO and CAISO experiences? One lesson is that it is certainly possible to integrate variable resources like wind into economic dispatch for an overall benefit to all parties involved. This integration works best when there is a large and well-functioning market and a variety of offtakers and market actors that are free to optimize for economics without too many contractual barriers. In MISO, the transition to the Dispatchable Intermittent Resource protocol has led to more efficient system operation, providing developers with better revenue predictability and lower risk.
Another lesson is that as variable resources deliver a larger share of total generation, addressing flexibility issues and other inefficiencies in the existing grid will become increasingly important. In California, developers, already concerned about the potential of future curtailment, have begun addressing this risk by adjusting contract structures to limit curtailment. Policymakers will need to think not only about how new entrants need to adapt, but also about how current practice, policies and systems adapted for incumbent stakeholders need to change to maximize the overall efficiency and value of the resources on the system.
Thank you to Jeff Bladen, Mark Ahlstrom, Jim Caldwell, Uday Varadarajan, Arno Harris, Michael Wheeler, and Andres Pacheco for their input on this piece. The authors are responsible for its final content.