Markets for solar PV and wind power are expanding rapidly in many regions of the world due to declining costs and to a variety of benefits and opportunities that these technologies can provide. Some countries already meet significant shares of their electricity demand with these variable renewable resources.
While power systems have always had to accommodate variability in both supply and demand, the growing adoption of variable renewable energy (VRE) is changing how power systems are planned, designed and operated. This is because the variability of output from solar and wind power means that more flexibility is required from the rest of the power system, including generating resources, distribution networks and even electricity consumers.
In areas where demand is growing (notably in developing economies), there is an opportunity for new and less-established power systems to grow in concert with higher shares of renewable generation as more flexible systems are developed. It is already possible to avoid lock-in of traditional “baseload” generation by using VRE to provide low-cost energy access and while avoiding costly investments in traditional, and less flexible, generation and grid infrastructure.
In all contexts, a shift away from the traditional “baseload thinking” in power system planning and operations will facilitate optimal integration of growing shares of VRE while providing on-demand, reliable and affordable electricity.
Power Systems: Traditional Design
Both traditional, centralised power systems and distributed, often renewable, energy systems strive to balance the supply and demand of electricity at all times. Their primary objective is to provide access to reliable electricity services at a reasonable price. Traditionally, centralised power systems use electric power facilities classified into three general, and sometimes overlapping, categoriesi:
Baseload generation – Generators such as coal, nuclear and large hydropower facilities are optimised for operation at full output with minimal interruption to meet the minimum level of load ii over a given period of time (days, weeks or months).The cost characteristics of traditional baseload generators can vary somewhat, but they typically have relatively high capital costs and relatively low variable costs. This means that these systems achieve their lowest average cost of energy if they are run continuously at full output. Baseload is usually considered an inflexible class of generation, meaning that output cannot be adjusted quickly up or down, with the exceptions of hydropower and geothermal power. The term baseload is an economic paradigm that has been in existence for many decades, but its usefulness is beginning to change in some regions, as explored below.
Intermediate or mid-merit generation – This includes natural gas combined-cycle generation and sometimes hydropower capacity that is able to adjust power output up or down in response to fluctuating demand.1 The generators supplement power that is provided by baseload generation. This class of generators is typically designed for frequent flexible operations and may be more expensive to operate than baseload because variable costs (e.g., fuel) may be higher, but also because all costs are spread out over fewer hours of the year.
Peaking generation – These are generators such as gas- or oil-fired turbines, or diesel generators, that are called on infrequently to meet peak load during periods of very high demand or extreme weather events. They also may be used when other generators or transmission lines are unavailable due to unforeseen outages. These generators are often relatively inefficient and the most expensive form of generation per unit of output, but they are used for short-term and incidental operation because their high variable costs are offset by low capital costs compared to plants optimised for full-time operation.
Demand has always been variable and to some degree unpredictable due to weather and uptake of emerging technologies. To a lesser degree, supply also has been variable given that generators or transmission lines can go offline unexpectedly, even in the most advanced power systems. In the face of this demand and supply variability, system operators have used flexible generation (and flexible demand to a lesser extent) to keep supply and demand in balance. In other words, large, generally inflexible baseload plants such as coal and nuclear have always been complemented by flexible generation in order to meet time-variable demand.
In countries with less mature power systems and/or rapidly growing economies, the demand for electricity may be more difficult to predict in advance because usage patterns are less established and consumers may tend to use more electricity as they add new electrical devices to their homes and businesses. Supply-side variability also may be more pronounced in such countries. Load shedding, or an interruption of energy supply to certain areas in response to balancing challenges, is more common in developing countries. In response, back-up generators are used frequently, and in some cases daily.2 Where reliable electricity infrastructure is lacking, introducing flexibility to enable higher shares of VRE can help alleviate pressure on strained power systems, and offer better service to customers as demand grows.
iGenerators also must supply “ancillary services” such as voltage support and various forms of reserve capacity to fine-tune the matching of supply and demand and to ensure reliability. For more on ancillary services, see, for example, Martin Beck and Marc Sherer, “Overview of Ancillary Services”, Swiss Grid, 4 December 2010, http://tinyurl.com/gu5zx4u, and Eric Ela, Michael Milligan and Brandon Kirby, Operating Reserves and Variable Generation (Golden, CO: National Renewable Energy Laboratory (NREL), 2011), http://www.nrel.gov/docs/fy11osti/51978.pdf.i
iiLoad in this context refers to the total amount of electricity demand from all industrial, commercial and residential sources at any given moment.ii
What Is Changing?
Around the world, markets for variable solar and wind power are expanding rapidly for a variety of reasons. These generation sources represent myriad benefits that set them apart from their traditional counterparts. For example, they draw on local resources, can be installed quickly in centralised or decentralised configurations, do not necessarily rely on existing infrastructure (and, unlike traditional systems, are not hampered by a lack of existing infrastructure), do not emit greenhouse gases or other pollutants during generation and generally require little water to operate. Due to their decentralised nature, they also may improve system security in the face of extreme events. In many regions of the world, VRE is now the lowest-cost source of newly constructed power generation available, thanks to rapidly declining capital costs and zero fuel costs.3
Subsequent to the growth of VRE in many locations, traditional baseload generators are beginning to lose their economic advantage and may no longer be the first to dispatch energyi. This means that once wind or solar power plants are put in service, all else being equal, it is most cost-effective to use all of the energy that they produce, within the bounds of system constraints, and as long as the additional system costsii are not excessive.
With VRE providing increasing amounts of first-in-line generation, several key aspects of power system operation and planning will change:
■ As the lowest marginal-cost form of energy on the system, VRE generation in most circumstances will be used when it is available, even if the next cheapest (in terms of marginal cost) generator must reduce its output.
■ In established power systems, the market share for traditional baseload generators as providers of bulk energy will decline as operators opt instead for least-cost VRE generation. This, in turn, will make near-constant operation less viable if all VRE is to be effectively utilised, further reducing the cost-competitiveness of baseload generation relative to VRE. Under certain circumstances, traditional baseload generators may begin to operate in a fashion similar to intermediate providers by ramping their output more frequently, to the extent that plant-specific economics and technical constraints allow, raising their average cost per unit of output.4
■ The remaining energy demand beyond that met by VRE (i.e., residual load) will be more variable, due to the impacts of variable wind and solar generation. Generators that must serve this more variable residual load will be required to operate more flexibly than under the old paradigm.5 (→ See Figure 59.)
In less-developed power systems, integrating flexibility into power system planning will enable higher shares of VRE up-front and reduce the need for traditional, near-constant, baseload operation.
iAs a result, traditional generators may face concerns of revenue sufficiency, or lost revenue, in systems that see growing shares of near-zero marginal cost VRE. See Bethany Frew et al., Revenue Sufficiency and Reliability in a Zero Marginal Cost Future (Golden, CO: NREL, 2016), http://www.nrel.gov/docs/fy17osti/66935.pdf.i
iiAdditional system costs may include balancing costs (adjustments of dispatchable power plants that respond to short-term variability of VRE), grid costs (that can include additional transmission) and costs related to any back-up capacity that may be required. Falko Ueckerdt et al., “System LCOE: What are the costs of variable renewables?” Energy, vol. 63 (15 December 2013), pp. 61–75, http://www.sciencedirect.com/science/article/pii/S0360544213009390. Such costs of integration are highly location-specific – they depend on available power system resources as well as on the characteristics and penetration levels of the specific VRE being used. D. Lew et al., The Western Wind and Solar Integration Study Phase 2 (Golden, CO: NREL, 2013), http://www.nrel.gov/docs/fy13osti/55588.pdf.ii
iiiDispatch intervals refer to the time between each new market auction. Shorter dispatch intervals allow dispatch to adjust to renewable variations more quickly and accurately, reducing balancing needs. See Eric Martinot, “Grid integration of renewable energy: flexibility, innovation and experience”, Annual Review of Environment and Resources, vol. 41 (2016), pp. 223-51, http://www.annualreviews.org/doi/abs/10.1146/annurev-environ-110615-085725.iii
ivA balancing area in this context refers to a system of power generation and transmission within the jurisdiction of a single authority.iv
Many technologies and approaches exist to increase flexibility on both the demand and supply sides of power generation.6 Options such as improved VRE forecasting, use of shorter system dispatch intervalsiii, co-ordination and trade of electricity supply across larger balancing areasiv and electricity storage can increase system flexibility.7 (→ See Storage section in Enabling Technologies chapter.) In many countries, grid operators also have used increasingly sophisticated forms of demand response, or incentives that influence customers to shift their use of power to minimise the cost of keeping supply and demand in balancei.8
Variable renewable energy systems themselves also can provide flexibility. Operators and regulators are increasingly requiring the use of VRE technology features that provide services to the grid.9 In Germany, for example, many solar PV systems are required to use smart inverters that ensure ongoing operation in the event of a grid disturbance.10 Characteristics of VRE power purchase agreements also are evolving in many settings to promote more flexible power systems and to limit curtailment of excess energy generated by VRE.11
Conventional generation and certain hydropower resources can be equipped with advanced technologies to provide additional flexibility in electricity supply. In Canada, for example, a coal generating station that originally was designed to provide baseload generation was successfully retrofitted to decrease minimum generation levels and to cycle on and off up to four times per day.12 Hydropower plants can incorporate variable speed technology, which increases flexibility by allowing power regulation in different modes of operation.13 One such plant began operation in India in 2016.14
The appropriate selection or mix of these flexibility options will depend on local circumstances. Ireland, for example, has limited opportunities for electricity trade, yet it relies on wind power for approximately one-quarter of its total electricity generation.15 Similarly, ERCOT, the power system operator in the US state of Texas, has very limited capability to import or export power to other interconnections, but generates far more wind energy than any other US state.16 Both Ireland and Texas rely on other sources of flexibility, including flexible generation, state-of-the-art wind forecasting and transmission expansion. Uruguay, which supplies 22% of its annual electricity with wind, relies on reservoir hydropower and interconnection with grids of neighbouring countries to provide flexibility.17 As the penetration of VRE increases, different power systems can employ a combination of flexibility options that are most appropriate and cost-effective under their different institutional, technological and economic contexts.
Based on different mixes of these flexibility mechanisms, VRE has already been integrated in 10 countries above double-digit shares of annual electricity generation without compromising the reliability of electricity supply.18 The ease of grid integration will vary from country to country.19 Typically, as the range of VRE penetration increases, so does the impact on power systems, requiring different prioritisation of response options to ensure adequate levels of flexibility.20 (→ See Table 4.)
Where electricity systems are developing, the most attractive option (in terms of both cost and practicality) may be to deploy infrastructure and operations with the flexibility necessary to handle high shares of VRE.
iThe simplest form of demand response is to shed load or to dictate when customers can consume. More advanced methods apply price incentives to encourage a shift of consumption to periods of relatively low demand.i
A New Planning Paradigm
In all contexts, power system planning plays a major role in setting the trajectory of electricity sector development. Where resources are strong, incorporating high shares of VRE alters planning in un-served and underserved areas because this removes constraints to build new generation capacity in geographic proximity to the existing power system; instead, new capacity can be placed where it makes sense to best serve new and existing customers. In such cases, distributed and VRE systems offer cost-competitive and often more immediate options for providing energy services.21
Traditional planning typically has been capacity-based, determining how many baseload, intermediate or peaking units are needed to meet projected energy demand in the future. As the penetration of higher shares of VRE increases, a different type of planning paradigm is required – one that takes into consideration the various costs and benefits derived from solar and wind power generation as well as the operational demands of VRE on system flexibility.22 In such a new, VRE planning paradigm, power system planners are able to identify the least-cost energy mix while maintaining the reliability of future energy supply.
Integrated Resource and Resiliency Planning (IRRP), sometimes known as Integrated Resource Planning, is a robust framework for identifying the optimal mix of where, how much and what types of power system resources will enable lowest-cost power sector development in the long term while also achieving goals related to reliability, climate, energy access and economic development.23 IRRPs are common among utilities in developed countries, and utilities in developing countries such as South Africa and Ghana are currently developing new IRRPs.24 IRRP modelling can integrate emerging best practices for demand response and for managing increasing shares of VRE, including high-quality representation of VRE resource potential, technical and financial implications of distributed VRE, transmission planning, and emerging technologies and operational practices for greater flexibility.25
The Ongoing Transition Away from Baseload
Countries in which high shares (20-40%) of VRE have been integrated (e.g., Denmark, Germany, Portugal, Uruguay and Cabo Verde) have demonstrated the shift away from the traditional baseload paradigm.26 In Denmark and Germany, interconnection with other European grids has helped to support peaks of 140% and 86.3%, respectively, of electricity generation from renewable energy.27 Cabo Verde, which supplies 25% of electricity with wind energy, plans to build an additional 20 MW of pumped storage capacity to help manage expanding renewable energy capacity on the island.28
Countries in which power demand is currently unmet or growing rapidly may face different conditions for integrating VRE into their power systems than developed countries, where demand for power is typically flat or declining. However, there may be administrative or institutional barriers that inhibit the development of more flexible systems. For example, power systems in developing countries may have baseload generators with mandated generation minimums and limited capital to expand transmission networks, and they may face decisions about extending the grid into new areas versus building mini-grids.29
Electric power facilities are long-term investments that involve complex supply chains and employ many people and therefore are subject to system inertia and related institutional, political and cultural barriers.30 Vested interests in the conventional baseload power system and lack of understanding of and education in new approaches and technological advancements are preventing many countries from moving towards higher shares of VRE, even when variable renewables might help reduce the overall cost of energy provision and improve the quality of energy services. Immature or poorly functioning institutions also can cause difficulties in both developed and developing countries, albeit to different extents.31
A range of planning, operational and institutional changes to the power system can be pursued to promote overall least-cost operation and investment strategies while preserving reliability.32 These strategies can also improve reliability and cost effectiveness in systems that are less developed, regardless of renewable energy penetration. As VRE resources and other enabling technologies – including storage, demand response and efficiency improvements – continue to achieve more favourable cost and performance characteristics, the incentive to deploy them will continue to increase, moving both new and existing power systems further from the baseload paradigm.
A) The Baseload Paradigm
In the early stages of progression to larger shares of variable renewable generation, power systems make some adjustments in their grid operations, develop forecasting systems for renewable energy production, and introduce improved control technology and operating procedures for efficient scheduling and dispatch.
B) The Early Transition
In the late stages of progression towards fully renewable power systems, variable renewable power will be integrated through advanced resource forecasting, grid reinforcements and strengthened interconnections, improved information and control technologies for grid operations, widespread deployment of storage technologies, greater efficiency and scope of demand response, and coupling of electricity, heating and cooling, and transport sectors.
C) A New Paradigm
- Hydropower can serve a variety of roles in power markets depending on resource endowments and physical topology, water management needs, and other social and environmental factors. In some countries, such as Brazil and China, hydropower can be primarily a baseload source of energy, while in others, hydropower may be restricted to providing load-following (mid-merit) functions. Still in others, pumped storage hydropower can function well as a source of peaking generation. For more, see Arun Kumar et al., “Chapter 5: Hydropower”, in Intergovernmental Panel on Climate Change (IPCC), Special Report on Renewable Energy Sources and Climate Change Mitigation (Cambridge, UK: 2011), .1
- Aaron Leopold, Power for All, personal communication with Renewable Energy Network for the 21st Century (REN21), 29 January 2017.2
- See, for example, Bloomberg New Energy Finance, Bloomberg New Energy Outlook 2016: Global Overview (London: 2016), , and International Energy Agency (IEA) Technology Collaboration Programme for Renewable Energy Technology Deployment (IEA-RETD), RETRANSITION – Transitioning to Policy Frameworks for Cost-Competitive Renewables (Paris: 2016), .3
- See, for example, IEA Clean Coal Centre, Increasing the Flexibility of Coal-fired Power Plants (London: 2014), .4
- Increasing flexibility entails specifically steeper ramping requirements, lower minimum turndown requirements and shorter lead-time for operating instructions. See, for example, Jaquelin Cochran et al., Flexibility in 21st Century Power Systems (Golden, CO: National Renewable Energy Laboratory (NREL), 2014), , and Aaron Bloom et al., Eastern Renewable Generation Integration Study (Golden, CO: NREL, 2016). Figure 59 based on IEA, Getting Wind and Sun onto the Grid - A Manual for Policy Makers (Paris: 2017), and on REN21, Renewables Global Futures Report – Great Debates Towards 100% Renewable Energy (Paris: 2017), . 5
- See, for example, Cochran et al., op. cit. note 5, and IEA, Flexible Power Systems to Integrate Large Shares of Renewables: The Value of Interconnections (Paris: 2014), 6
- Cochran et al., op. cit. note 5.7
- Paul Denholm, “The Role of Storage and Demand Response”, Greening the Grid, 2015, .8
- Greening the Grid, “Ancillary services”, .9
- Smart inverters are those that include digital architecture and bi-directional communications capabilities. See, for example, NREL, Advanced Inverter Functions to Support High Levels of Distributed Solar (Golden, CO: 2015), , and A.S. Awad, M. Abdel-Rahman and M. Abdel Latif Badr, “Low-voltage ride-through capability characterization of wind farms”, Electric Power Components and Systems”, vol. 40, no. 16 (2012), pp. 1808–19, .10
- Lori Bird, Jaquelin Cochran and Xi Wang, Wind and Solar Energy Curtailment: Experience and Practices in the United States (Golden, CO: NREL, 2014), .11
- Jaquelin Cochran, Debra Lew and Nikhil Kumar, Flexible Coal: Evolution from Baseload to Peaking Plant, 21st Century Power Partnership, 2013, .12
- Suzanne Pritchard, “The future for hydro – roundtable discussion”, International Water Power & Dam Construction, 2 March 2015, .13
- Power-technology.com, “Tehri Pumped Storage Plant, India”, , viewed 28 March 2017.14
- Eoin Burke Kennedy, “Over 23% of electricity demand now supplied by wind”, Irish Times, 23 December 2015, .15
- Brattle Group, Integrating Renewable Energy into the Electricity Grid: Case Studies Showing How System Operators Are Maintaining Reliability (Cambridge, MA: 2015), .16
- Uruguay Secretary of Energy, Ministry of Industry, Energy and Mining, Balance Energético Preliminar 2016 (Montevideo: 2017), ; Jonathan Watts, “Uruguay makes dramatic shift to nearly 95% electricity from clean energy”, The Guardian (UK), 3 December 2015, ; Uruguay Secretary of Energy, Ministry of Industry, Energy and Mining, personal communication with REN21, May 18, 2017. 17
- IEA, op. cit. note 5.18
- International Renewable Energy Agency (IRENA), REthinking Energy 2017: Accelerating the Global Energy Transformation (Abu Dhabi: 2017), p. 75, .19
- IEA, op. cit. note 5; IRENA, op. cit. note 19, p. 75. Table 4 based on IEA, op. cit. note 5, and on REN21, op. cit. note 5.20
- Practical Action, Poor People’s Energy Outlook 2016 (Bourton on Dunsmore, Rugby, Warwickshire, UK: 2016), .21
- See, for example, IRENA, Planning for the Renewable Future: Long-term Modeling and Tools to Expand Variable Renewable Power in Emerging Economies (Abu Dhabi: 2017), , and Jessica Katz, “The Evolution of Power System Planning with High Levels of Variable Renewable Energy”, Greening the Grid, 2016, .22
- Rachael Wilson and Bruce Biewald, Best Practices in Electric Utility Integrated Resource Planning (Montpelier, VT: RAP, 2013), .23
- Department of Energy, Republic of South Africa, “Integrated Resource Plan”, ; Obed Attah Yeboah, “Power Sector Master Plan in the Works”, B&FT Online, 16 September 2016, ; for an example of an IRP in a developed country, see, for example, PacifiCorp, “Integrated Resource Plan”, .24
- See, for example, Greening the Grid, “Renewable Energy Zones: Delivering Clean Power to Meet Demand”, 2016, .25
- Arthur Neslen, “Portugal runs for four days straight on renewable energy alone”, The Guardian (UK), 18 May 2016, ; Uruguay Secretary of Energy, Ministry of Industry, Energy and Mining, op. cit. note 17; Adam Withnall, “Cape Verde: The African country that plans to run on 100% renewable energy by 2020”, The Independent (UK), 29 September 2016, .26
- Arthur Neslen, “Wind power generates 140% of Denmark’s electricity demand”, The Guardian (UK), 10 July 2015, ; Agora Energiewende, Die Energiewende im Stromsektor: Stand der Dinge 2016 (Berlin: January 2017), p. 7, .27
- Withnall, op. cit. note 26; IRENA, Renewable Islands: Settings for Success (Abu Dhabi: 2014), p. 8, .28
- Kathryn M. O’Neill, “Going off grid: Tata researchers tackle rural electrification”, Massachusetts Institute of Technology, 21 January 2016, .29
- W. Brian Arthur, “Competing technologies, increasing returns and lock-in by historical events”, Economic Journal, vol. 99, no. 394 (1989), pp. 116-31, ; Gregory Unruh, “Understanding carbon lock-in”, Energy Policy, vol. 28, no. 12 (2000), .30
- See, for example, IRENA, REmap: Roadmap for a Renewable Energy Future (Abu Dhabi: 2016), , and Jaquelin Cochran et al., Integrating Variable Renewable Energy in Electric Power Markets: Best Practices from International Experience (Golden, CO: NREL, 2012), .31
- Additional information on these strategies and best practices can be found in Cochran et al., op. cit. note 31; IRENA, op. cit. note 22; and IEA, op. cit. note 5. See also Debra Lew et al., Wind and Solar Curtailment (Golden, CO: NREL, 2013), .32