Abstract
The electric power industry is undergoing dramatic change driven by rapid technological transformation, concern over climate change, and evolving market structures. As a result, there is a need for new models to help the industry better utilize resources in a time of increasing uncertainty and to help government policy makers better understand the impacts of their regulatory decisions. We provide a structured review of the operations research and management science literatures to describe the current operational and policy issues in the electric power industry, with a particular focus on issues surrounding electricity market design, renewable integration, effects of climate policy on electric power infrastructure, rise of electric powered vehicles, energy storage, and the growing interdependence between natural gas and electric power sectors. We identify the current research frontier and classify the existing research into clusters with respect to managerial issues addressed. We offer a forecast for where the electric power industry is going and describe some important public policy issues. Finally, we highlight research opportunities and discuss how the management science community can contribute by considering the interaction between operational considerations and electric power policy.
Introduction
The sources of and efficiency with which society uses energy is fundamental and of the utmost importance along a surprisingly large number of dimensions. Manufacturing competitiveness, economic growth, foreign policy, and environmental policy are just a few sectors that are directly affected by energy decisions. In recent years, the energy industry in general and the electric power industry, in particular, have been subject to rapid technological change that has had major impacts. For example, natural gas fracturing improved so rapidly from 2005 to 2010 that the impact of cheaper natural gas has been felt across the economy in sectors such as chemicals, manufacturing, residential heating, and electric power. The widely debated “war on coal” was arguably much more about the hydraulic fracturing driven price reduction of natural gas than explicit regulatory policies related to coal.
In addition to a revolution in the cost of extracting hydrocarbons, the industry is also on the cusp of a revolution in the way that electric power is generated and consumed. Two forces are driving change in this sector. The first is the rapid decrease in the price of renewable generation technology, primarily solar and wind that have generated such demand anomalies as the “California Duck” curve where net electricity demanded from the central grid falls at peak solar production hours and then rises rapidly when the sun sets. The second force is the dramatic reduction in the cost of information and computing technology which has the promise of making electric power demand far more responsive to price and other control signals. Indeed, in a reversal of over one hundred years of practice in ramping dispatchable supply to meet demand, we may be entering an era where electricity supply is the random variable that is addressed by adjusting demand to match supply. Or, more realistically, we are entering an era where both supply and demand have stochastic and controllable components, adding to the complexity of ensuring reliable electric service.
The rapid change on the supply side of the market, for both conventional and renewable technology and rapid change on the demand side because of information technology and the potential for improved storage have combined to create a need for new models to help the electric power industry better utilize and control energy and to help government policy makers better understand the impacts of regulatory decisions. We believe that the operations research and management science (OR/MS) community is in a strong position to provide the industry with valuable decision‐making support.
In this study, we review the OR/MS literature to describe current operational and policy issues in the electric power industry with the ultimate goal of encouraging management science scholars to do more research in the area. We believe there is great potential to do rigorous research that also has real impact on a critical industry. One criticism of OR/MS research is that real life details are often abstracted away in the pursuit of creating elegant models. Our hope is to encourage more research that retains the necessary detail to be relevant in an industry context. OR/MS scholars have the technical capability to solve problems at the intersection of business, economics, and policy. Large literatures on electric power industry exist at both the technical and policy levels, but OR/MS scholars are particularly well placed to bridge these focal areas. Given the long history of regulation and public investment, the potential for dual‐causality between operations and public policy decisions (Joglekar et al. 2016) is especially strong in the electric power industry. One motivation for this study is, therefore, to establish a roadmap for future OR/MS research that will take into account the mutually interacting dynamics between operations and public policy decisions in the electric power industry.
Our intended contribution is twofold. First, we provide a structured review of electric power industry research in the last 20 years (mostly within OR/MS journals) with citations to approximately 500 papers. We focus on topics that we find to be especially relevant to current operational and policy challenges, such as electricity market design, renewable integration, hydropower operations, energy storage, climate policy, risk management in the electric power industry, electrification of the transportation sector, and the growing interdependence between natural gas and electric power generation. Although this has been a major undertaking, we still go over only a subset of electric power industry research in general and, for feasibility, we completely omit related upstream industries such as oil and gas exploration and production. We hope that the survey and references can help researchers to locate papers in relevant areas. Second, and more importantly, we identify research frontiers in electric power industry research, classify the existing research into clusters with respect to managerial issues addressed, and discuss possibilities for future work that would extend this research frontier in directions with significant impact on practice.
We note that many papers included in this survey, particularly those on renewable energy, alternative fuel vehicles, or carbon policy can also be categorized under the umbrella of sustainable operations (Kleindorfer et al. 2013). The sustainable operations literature covers a wide range of issues including green product design, lean operations, green operations, and closed‐loop supply chains. We refer the readers to survey papers in this area for further information on opportunities for research regarding sustainable operations that are not covered in this paper (e.g., Corbett and Klassen 2006, Kleindorfer et al. 2013, Linton et al. 2007, Tang and Zhou 2012).
The rest of the study is organized as follows. Section 2 briefly describes the research methodology used in choosing and categorizing papers. Section 3 discusses the current frontier in the electric power industry research by OR/MS scholars with subsection 3.1 providing the background information on the industry and subsections 3.2–3.9 focusing on electricity market design, renewable integration, risk management in electricity markets, climate policy, electricity storage, hydropower operations, adoption and integration of plug‐in electric and hybrid vehicles, and natural gas industry and its impact on electric power generation, respectively. Section 4 summarizes the most promising research areas for OR/MS scholars and concludes the paper.
Methodology
Research on the electric power industry is vast and rapidly growing. Thus, our approach is informed by previous survey papers (Anderson and Parker 2013b, Krishnan and Ulrich 2001) particularly with respect to the journals sampled and the need to narrow scope to yield a manageable task. There are numerous academic journals dedicated to the energy industry (e.g.,
Specifically, we thoroughly reviewed the following journals spanning the years 1998–2018: 1)
Next, we clustered majority of these papers into the respective research areas discussed in sections 3.2–3.9. For each research area, we further clustered the papers with respect to managerial issues addressed, which are summarized in Tables 2–10. Papers that are related to the electric power industry, but that do not fall under the main research areas in sections 3.2–3.9 are sampled and catalogued in Table 1. In each section, we discuss only a small subset of the references, but we hope that Tables 1–10 stand alone as a guide to articles that are likely to be most useful to readers.
A Sample of Electric Power Industry Research in OR/MS Journals
| Research topics | Selected references |
|---|---|
| Unit commitment/Optimal dispatch models: Optimal schedule and dispatch of power generation resources | Takriti and Krasenbrink (1999), Nowak and Römisch (2000), Takriti et al. (2000), Takriti and Birge (2000), Doorman and Nygreen (2003), Viana et al. (2003), Valenzuela and Mazumdar (2003), Thompson et al. (2004), Sen et al. (2006), Dahal et al. (2007), Troncoso et al. (2008), Cerisola et al. (2009), Toczy lowski and Zoltowska (2009), Corchero and Heredia (2011), Kjeldsen and Chiarandini (2012), Murillo‐Sánchez et al. (2013), Papavasiliou and Oren (2013), Zugno and Conejo (2015), Lorca et al. (2016), van Ackooij and Malick (2016), Zugno et al. (2016), and Morales and Pineda (2017) |
| Valuation and/or optimal investment level for generation, storage, or transmission assets: single‐firm level decision making, includes real options‐based papers | Tseng and Barz (2002), Takizawa and Suzuki (2004), Thompson et al. (2004), Tseng and Lin (2007), Boomsma et al. (2012), Thompson (2013), Farzan et al. (2015), Hu et al. (2015), Bruno et al. (2016), Drake et al. (2016), Hach et al. (2016), Kettunen and Bunn (2016), and Welling (2016) |
| Optimal capacity expansion for transmission networks and generation assets: long‐term decisions by central planners | Gardner and Rogers (1999), Singh et al. (2009), Villumsen and Philpott (2012), Feng and Ryan (2013), ONeill et al. (2013), Parpas and Webster (2014), Huppmann and Egerer (2015), Jenabi et al. (2015), Sauma et al. (2015), Tolis (2015), Grimm et al. (2016), Georgiou (2016), Munoz et al. (2016), Pineda et al. (2016), and Pineda and Morales (2016) |
| Optimal operation of generation assets | Nowak and Römisch (2000), Mijangos (2005), Makkonen and Lahdelma (2006), Rong and Lahdelma (2007), Cirre et al. (2009), Bosman et al. (2012), Kalczynski (2012), Zhang et al. (2013), Jochem et al. (2015), and Steffen and Weber (2016) |
| Decision support systems for power market policy making and operations | Georgopoulou et al. (1998), Bergey et al. (2003), Papadopoulos and Karagiannidis (2008), Hunt et al. (2013), Mattiussi et al. (2014), Bertsch and Fichtner (2015), and Pinto et al. (2015) |
Electricity Markets
| Research topics | Selected references |
|---|---|
| Finding market equilibria in deregulated electricity markets (Cournot, Bertrand, Stackelberg, or Supply Function Equilibria for spot and forward prices and generation levels) | Jing‐Yuan and Smeers (1999), Younes and Ilic (1999), Lavigne et al. (2000), Boucher and Smeers (2001), Schuler (2001), Bessembinder and Lemmon (2002), Bunn and Oliveira (2003), Bushnell (2003), Garcia et al. (2005), Rudkevich (2005), Wu and Kleindorfer (2005), de Haro et al. (2007), Hobbs and Pang (2007), Hu and Ralph (2007), Wang et al. (2007), 2007,2008Yao et al.), Anderson and Hu (2008), Bushnell et al. (2008), Wilson (2008), Holmberg (2009), Shanbhag et al. (2011), Thompson (2013), Filomena et al. (2014), Peura and Bunn (2015), and Ruddell et al. (2016) |
| Examination of market power or implicit collusion | Schuler (2001), Borenstein et al. (2002), Bunn and Oliveira (2003), Bushnell (2003), Murphy and Smeers (2005), Bushnell et al. (2008), Anderson and Cau (2009), Banal‐Estañol and Micola (2009), Murphy and Smeers (2010), Anderson and Cau (2011), Oh and Thomas (2013), and Peura and Bunn (2015) |
| Capacity expansion in electricity markets | Murphy and Smeers (2005), Wu et al. (2005), Bunn and Oliveira (2008), Murphy and Smeers (2010), Ehrenmann and Smeers (2011), Pineau et al. (2011), Filomena et al. (2014), Bunn and Oliveira (2016), and Oliveira and Costa (2018) |
| Effect of transmission costs and constraints on market behavior | Younes and Ilic (1999), Kattuman et al. (2004), Daxhelet and Smeers (2007), Contreras et al. (2009), Liu and Nagurney (2009), Downward et al. (2010), Ruddell et al. (2016), and Holmberg and Philpott (2018) |
| Electricity pricing in markets with nonconvexities | O'Neill et al. (2005), Bjørndal and Jornsten (2008), Araoz and Jörnsten (2011), Liberopoulos and Andrianesis (2016), and Fuller and Celebi (2017) |
| Optimal level of generation to offer to the market (i.e., optimal bids) | Anderson and Philpott (2002), Neame et al. (2003), Pritchard and Zakeri (2003), Kian and Cruz (2005), Triki et al. (2005), Fleten and Kristoffersen (2007), Aparicio et al. (2008), Beraldi et al. (2008), Corchero and Heredia (2011), Kim and Powell (2011), Boomsma et al. (2014), and Steeger and Rebennack (2017) |
| Bidding behavior, auction design and implementation | Mount (2001), Singer (2002), Schummer and Vohra (2003), Elmaghraby (2005), Hortacsu and Puller (2008), Meeus et al. (2009), Morales et al. (2014), Fabra and Reguant (2014), Reguant (2014), Derinkuyu (2015), and Madani and Van Vyve (2015, 2018) |
| Electricity market design and its effect on economic and environmental outcomes | Lavigne et al. (2000), Kleindorfer et al. (2001), Cramton and Stoft (2005), Bunn and Oliveira (2008), Milligan and Porter (2008), Ehrenmann and Neuhoff (2009), Zhao et al. (2010), Chen et al. (2011), Ehrenmann and Smeers (2011), Muratore (2011), Cramton et al. (2013), Rosen and Madlener (2013), Morales et al. (2014), Franco et al. (2015), Jiang et al. (2016), Hach et al. (2016), Kök et al. (2016), Pineda et al. (2016), Ritzenhofen et al. (2016), and Siddiqui et al. (2016) |
| Electricity tariffs and demand response | Oren (2001), Kamat and Oren (2002), Pettersen et al. (2005), Baldick et al. (2006), Sueyoshi and Tadiparthi (2008), Hatami et al. (2009), Ding et al. (2012), Latifoğlu et al. (2013), Puller and West (2013), Zakeri et al. (2014), Tsitsiklis and Xu (2015), Ata et al. (2016), Kök et al. (2016), Lohmann and Rebennack (2016), Ströhle and Flath (2016), Valogianni and Ketter (2016), Dong et al. (2017), Garttner et al. (2018), and Schlereth et al. (2018) |
| Net metering | Darghouth et al. (2011), Cai et al. (2013), Borlick and Wood (2014), Costello and Hemphill (2014), Eid et al. (2014), Hu et al. (2015), Satchwell et al. (2015), Darghouth et al. (2016), Castaneda et al. (2017), and Gagnon et al. (2017) |
Electricity Markets Open Questions
| Electricity markets open research questions | |
|---|---|
| Operational | Policy |
| What is the potential impact of dynamic electricity pricing on capacity investments, demand response adoption, emission levels, and technology mix of electricity generation portfolios? | How can a platform market for the electric power industry be developed to improve the integration of distributed energy resources into electricity distribution systems? |
| How do different rate structures for net metering affect renewable energy capacity investments? | During the platformization of the grid, how should utility business models be revised to avoid the utility death spiral? |
| How do different electricity tariff structures affect demand response? | How can the missing money problem be solved such that the market provides the necessary incentives to build adequate generation capacity? Can “energy‐only” designs be sufficient or are capacity markets a necessary part of the design? |
| How should unit commitment models and market clearing algorithms be adjusted in European markets to better handle the market coupling among interconnected power systems? How does large‐scale integration of renewables factor into these adjustments? | What kind of electricity tariffs would be ideal to reflect the varying cost of electricity? |
Renewable Integration
| Research topics | Selected references |
|---|---|
| Unit commitment and optimal dispatch: The impact of large scale wind integration on optimal schedule and dispatch of power generation resources | Bouffard and Galiana (2008), Wang et al. (2008), Ruiz et al. (2009), Sioshansi and Short (2009), Tuohy et al. (2009), Constantinescu et al. (2011), Papavasiliou and Oren (2013), Wu and Kapiscinski (2013), Petersen et al. (2016), and Schulze et al. (2017) |
| Optimal capital investment into a renewable generation asset | Murphy and Smeers (2010), Ehrenmann and Smeers (2011), Muñoz et al. (2011), Boomsma et al. (2012), Hu et al. (2015), Shittu et al. (2015), Alizamir et al. (2016), Bonneuil and Boucekkine (2016), Bruno et al. (2016), Wang et al. (2016), Welling (2016), and Aflaki and Netessine (2017) |
| Optimal level of generation to bid in the forward market | Pritchard and Zakeri (2003), Morales et al. (2010), Dent et al. (2011), Kim and Powell (2011), Bitar et al. (2012), and Botterud et al. (2012) |
| Joint operation of an intermittent power generation and electricity storage | Korpaas et al. (2003), Castronuovo and Lopes (2004), Brunetta and Tina (2007), Brown et al. (2008), Garcia‐Gonzalez et al. (2008), Kim and Powell (2011), Mauch et al. (2012), Kuznia et al. (2013), Jiang and Powell (2015a), Billionnet et al. (2016), Moarefdoost and Snyder (2015), Qi et al. (2015), Velik and Nicolay (2016), and Kamdem and Shittu (2017) |
| Effect of transmission constraints on renewable integration | Sioshansi and Short (2009), Papavasiliou and Oren (2013), Fischlein et al. (2013), and Qi et al. (2015) |
| Electricity markets and renewables: Market design issues that needs to be resolved to facilitate renewable integration; policy interventions to support renewable integration; the effect of renewable policy interventions on electricity markets | Cramton and Stoft (2005), Milligan and Porter (2008), Ehrenmann and Smeers (2011), Cramton et al. (2013), Morales et al. (2014), Franco et al. (2015), Jiang et al. (2016), Hach et al. (2016), Ritzenhofen et al. (2016), Siddiqui et al. (2016), Al‐Gwaiz et al. (2017), and Pineda et al. (2018) |
| Optimal design of energy systems with intermittent generation (e.g. optimal energy mix, optimal wind or solar farm placement, optimal wind farm portfolios, optimal design of hybrid power systems and microgrids for isolated communities) | Ferrer‐Marti et al. (2013), Kuznia et al. (2013), Le Cadre et al. (2015), Qu et al. (2015), Thomaidis et al. (2016), Billionnet et al. (2016), Triadó‐Aymerich et al. (2016), Zografidou et al. (2016), and Kamdem and Shittu (2017) |
| Optimizing the operations of intermittent generationassets | Zhang et al. (2013), Lima et al. (2015), Steffen and Weber (2016) |
| Open research questions | |
|---|---|
| Operational | Policy |
| What is the impact of large scale renewable integration on optimal schedule and dispatch of power generation resources? How is that impact different in European markets compared to that in US markets considering the market coupling among interconnected power systems in Europe? | What market policies or regulations can help improve the large‐scale integration of renewables? How do these policies differ in nodal (e.g., U.S.) versus zonal (e.g., European) markets? |
| What is the effect of transmission constraints on renewable integration? | Is the large‐scale integration of renewables more efficient in a regulated or a deregulated electricity market? |
| How should hybrid power systems, zero‐energy buildings, and microgrids that rely on distributed energy resources be designed? | What should be the public incentive policies to disseminate distributed energy resources? |
| What is the optimal bidding strategy for wind and solar power? How does the bidding strategy change if the generating unit has dedicated electricity storage? | What are the states’ optimal incentive policies for generation capacity expansion to reach renewable portfolio targets? |
Risk Management in the Electric Power Industry
| Research topics | Selected references |
|---|---|
| Electricity forward and futures pricing | Bessembinder and Lemmon (2002), Lucia and Schwartz (2002), Fleten and Lemming (2003), Longstaff and Wang (2004), Bunn (2004), Carmona and Coulon (2014), Islyaev and Date (2015), and Caldana et al. (2017) |
| Pricing of electricity contracts and derivatives | Kwon et al. (2006), Thompson (2013), Islyaev and Date (2015), and Wu and Babich (2012) |
| Electricity trading through forward and spot markets | Sen et al. (2006), Kwon et al. (2006), Dong and Liu (2007), Falbo et al. (2010), Oliveira et al. (2013), and Mari et al. (2017) |
| Risk management for load serving entities | Anderson and Hu (2008), Oum et al. (2006), Hatami et al. (2009), Oum and Oren (2009, 2010), Aïd et al. (2011), Boroumand and Zachmann (2012), Boroumand et al. (2015), and Downward et al. (2016) |
| Financial trading of electricity through electricity derivatives | Deng et al. (2001), Vehviläinen and Keppo (2003), Kleindorfer and Li (2005), Deng and Oren (2006), and Doege et al. (2009) |
| Review of electricity risk management practices | Sioshansi (2002), Eydeland and Wolyniec (2003), Deng and Oren (2006), and Liu et al. (2006) |
| Open research questions | |
|---|---|
| Operational | Policy |
| What are the best market and non‐market mechanisms of risk management for load serving entities? | How can the electric power industry develop an integrated and comprehensive risk management function that encompasses not just market risks but also policy risks? |
| How can the electric power industry better integrate financial and operational hedging for improved risk management? | What are the impacts of regulation on financial assessment of operational risks and insuring against these risks? |
| What will be the impact of improved storage technology on financial and operational hedging? | |
Climate Policy and Its Impact on the Electric Power Industry
| Research topics | Selected references |
|---|---|
| Climate policy design and implementation | Kanudia and Loulou (1998), Jaehn and Letmathe (2010), Zhao et al. (2010), Carmona and Hinz (2011), Chen et al. (2011), İşlegen and Reichelstein (2011), Caro et al. (2013), Nyiwul et al. (2015), Sunar and Plambeck (2016), İşlegen et al. (2016), Siddiqui et al. (2016), Drake et al. (2016), Chen and Kettunen (2017), and Drake (2018) |
| Comparison of climate policy alternatives | Fan et al. (2010), Zhao et al. (2010), Chen et al. (2011), He et al. (2012), Nyiwul et al. (2015), Drake et al. (2016), İşlegen et al. (2016), and Chen and Kettunen (2017) |
| Effect of climate policy on generation investment and technology choice | Fan et al. (2010), Kettunen et al. (2011), Drake et al. (2016), Kettunen and Bunn (2016), Chen and Kettunen (2017), and Aflaki and Netessine (2017) |
| Effect of climate policy on production, procurement, facility location and supply chain decisions | Gong and Zhou (2013), Krass et al. (2013), Cachon (2014), Park et al. (2015), İşlegen et al. (2016), and Yuan et al. (2018) |
| Firm level compliance strategies and voluntary emissions reduction | Subramanian et al. (2007), Kroes et al. (2012), Jira and Toffel (2013), and Jacobs (2014) |
| Carbon leakage and border adjustments | Condon and Ignaciuk (2013), Chen et al. (2011), İşlegen et al. (2016), Drake (2018), and Sunar and Plambeck (2016) |
| Effect of climate policy on R&D investment and technology portfolios | Baker and Shittu (2006, 2008), Bosetti and Tavoni (2009), Baker and Solak (2011, 2014), and Shittu (2014) |
| Open research questions | |
|---|---|
| Operational | Policy |
| How can operations and supply chain management practices be revised to reduce emissions? How can OR/MS research help firms strike a balance between sustainability and profitability? | Can climate policies be improved by bringing in an operations management perspective, such as incorporating firm level capacity, production, and emission abatement decisions into policy models? |
| How does uncertainty in climate policy affect capacity investment, technology choice, and production decisions? | How should uncertainty in technological learning rates be factored in when designing climate policy? |
| How should uncertainty in technological learning rates be factored in generation planning models? | How should policy address “carbon leakage” challenges? |
Electricity Storage
| Sample literature | |
|---|---|
| Research topics | Selected references |
| Arbitrage value of storage | Graves et al. (1999), Walawalkar et al. (2007), Figueiredo et al. (2006), and Sioshansi et al. (2009, 2011) |
| Optimal operating (e.g., charging, discharging, bidding) policy for the storage | Brown et al. (2008), Harsha and Dahleh (2015), Löhndorf et al. (2013), van de Ven et al. (2013), Zhou et al. (2015), Granado et al. (2007), Xi et al. (2014), Paine et al. (2014), Jiang and Powell (2015a,b), and Moarefdoost and Snyder (2015) |
| Optimal storage capacity | Harsha and Dahleh (2015), Brown et al. (2008), Brunetta and Tina (2007), Ru et al. (2013), Babrowski et al. (2015), Jiang et al. (2016), Billionnet et al. (2016), and Qi et al. (2015) |
| Effects of integrating electricity storage on social welfare | Sioshansi (2010, 2011b), Schill and Kempert (2011), and Sioshansi (2014) |
| Effects of integrating electricity storage on emissions | Sioshansi (2011a), Greenblatt et al. (2007), and Denholm et al. (2005) |
| Value of storage based on multiple types of service that the storage can provide | Walawalkar et al. (2007), Drury et al. (2011), and Xi et al. (2014) |
| Open research questions | |
|---|---|
| Operational | Policy |
| How can multiple uses of electricity storage that may be complements or substitutes be incorporated into storage valuation models? | How should the electricity markets be redesigned to enable widespread integration of electricity storage? |
| How can multiple uses of electricity storage that may be complements or substitutes be incorporated when optimizing storage operations? | Is energy storage integration more efficient in a regulated utility or deregulated electricity market? |
| What are the optimal infrastructure combinations of electricity storage with traditional (hydro) and intermittent (solar and wind) power generation methods? | How will commodity market regulations adapt to changes in storage capabilities? |
Hydropower Operations
| Sample literature | |
|---|---|
| Research topics | Selected references |
| Optimal operation of hydropower plants | Edirisinghe et al. (2000), Escudero (2000), Nowak and Römisch (2000), Philpott et al. (2000), Cai et al. (2001), Pritchard and Zakeri (2003), Lino et al. (2003), Castro and González (2004), García‐González et al. (2007), Latorre et al. (2007), Fleten and Kristoffersen (2008), Stamford Da Silva and Campello De Souza (2008), Azevedo et al. (2009), De Ladurantaye et al. (2009), Frangioni and Gentile (2009), Cerisola et al. (2012), Guigues and Sagastizábal (2012), Denault et al. (2013), Densing (2013), Pritchard (2015), Ghaddar et al. (2015), Vojvodic et al. (2016), Séguin et al. (2017), Zephyr et al. (2017), Carpentier et al. (2018), and Gauvin et al. (2018) |
| Wind and hydro coordination | Belanger and Gagnon (2002), Castronuovo and Lopes (2004), Jaramillo et al. (2004), Bueno and Carta (2006), Benitez et al. (2008), Vespucci et al. (2012), Löhndorf et al. (2013), and Steffen and Weber (2016) |
| Investment in hydropower assets | Chaton and Doucet (2003), Thompson et al. (2004), Bøckman et al. (2008), and Bruno et al. (2016) |
| Open research questions | |
|---|---|
| Operational | Policy |
| How should wind and hydropower operations be coordinated? How does hydropower's new role of balancing the fluctuations in wind output interact with its other functions such as irrigation? | Given hydropower's new role of balancing the fluctuations in wind output, should new incentives be designed to promote investments into hydropower? |
| How does the uncertainty introduced by climate change affect hydropower operations? | Given the environmental consequences of building new dams, how should policy makers reconcile the needs of various stakeholders in designing hydropower incentives and regulations? |
| How does the uncertainty introduced by climate change affect hydropower capacity investments? | |
| How does the uncertainty introduced by climate change influence the effectiveness of hydropower in achieving climate targets? For example, given the increased risk of prolonged draughts in some regions due to climate change and that the resulting gap in generation tends to be picked up by fossil‐fuel‐based generators, what is the net effect of hydropower capacity expansions on achieving climate targets? | |
Plug‐in Electric and Hybrid Vehicles
| Research topics | Selected references |
|---|---|
| Impact of integrating PEVs into the power system | Parks et al. (2007), Lemoine et al. (2008), Samaras and Meisterling (2008), Stephan and Sullivan (2008), Sioshansi and Denholm (2009), Sioshansi et al. (2010), Wang et al. (2010), Sioshansi (2012), and Kahlen et al. (2018) |
| Market diffusion of electric vehicles: Barriers in front of the mass adoption | Struben and Sterman (2008), Luo et al. (2014), Cohen et al. (2015), Lim et al. (2015), Gnann and Plotz (2015), He et al. (2017), Kuppusamy et al. (2017), and Shao et al. (2017) |
| Optimal deployment and operation of refueling infrastructure | Capar et al. (2015), Lim and Kuby (2010), Raviv (2012), He et al. (2013), Mak et al. (2013), Nurre et al. (2014), Avci et al. (2015), Chung and Kwon (2015), Hung and Michailidis (2015), Yang and Sun (2015), Zhao and Ma (2016), and Yildiz et al. (2016) |
| Vehicle routing and scheduling problem for electric (and other alternative fuel) vehicles with limited ranges | Boyaci et al. (2015), Goeke and Scheider (2015), Hung and Michailidis (2015), Doppstadt et al. (2016), Hiermann et al. (2016), Pimenta et al. (2016), Wen et al. (2016), and Schiffer and Walther (2017) |
| Open research questions | |
|---|---|
| Operational | Policy |
| How do different types of refueling infrastructure compare in terms of their effectiveness of promoting the adoption of plug‐in electric or hybrid vehicles (PEVs)? | What kind of policies would help the adoption of PEVs the most? Should governments subsidize consumers to promote early adoption of electric vehicles, or the private sector to encourage investment in refueling infrastructure, or a mix of the two? |
| What would be the best way to incorporate demand uncertainty in PEV adoption and refueling infrastructure models? | What are some policies and market designs that would incentivize desirable charging patterns for PEVs to reduce their impact on the grid in terms of capacity requirements? |
| What would be the effect of wide‐spread availability of V2G services on power grid performance? | What kind of market rules and policies would facilitate the integration of V2G services? How would those policies differ in a regulated utility versus deregulated market? |
| What would be the potential capacity and infrastructure requirements for V2G integration? | How does the volatility in fossil fuel prices affect the adoption of electric vehicles and the effectiveness of policies that support adoption, such as government subsidies? |
| How would revenue streams from V2G affect consumer adoption of PEVs and vice versa? | Should recharge network operators be required to open access to their networks? |
Natural Gas Industry
| Research topics | Selected references |
|---|---|
| Coordinated operation of the electric power and natural gas industries | Li et al. (2008), Liu et al. (2009), Hibbard and Schatzki (2012), Alabdulwahab et al. (2015), Bai et al. (2016), Cui et al. (2016), and Chiang and Zavala (2016) |
| Coordinated expansion planning of the electric power and natural gas industries | Chaudry et al. (2008), Unsihuay‐Vila et al. (2010), Chaudry et al. (2014), and Zhang et al. (2015) |
| Natural gas and LNG storage optimization and valuation | Butler and Dyer (1999), Gray and Khandelwal (2004), Carmona and Ludkovski (2010), Chen and Forsyth (2007), Boogert and De Jong (2008), Thompson et al. (2009), Secomandi (2010a, 2010b), Lai et al. (2010, 2011), Felix and Weber (2012), Wu et al. (2012), Secomandi et al. (2015), and Nadarajah et al. (2015) |
| Optimal procurement and/or inventory policy for natural gas and other commodities | Guldmann and Wang (1999), Jaillet et al. (2004), Mendelson and Tunca (2007), Muthuraman et al. (2008), Aouam et al. (2010), Wu and Chen (2010), Devalkar et al. (2011), Goel and Gutierrez (2011), Chen et al. (2013), Kouvelis et al. (2013), Secomandi and Kekre (2014), Chen et al. (2015), and Aouam et al. (2016) |
| Equilibrium models of natural gas markets (looking at price formation, market power, capacity adequacy, etc.) | Gabriel et al. (2001), Gailly et al. (2001), Gabriel et al. (2005a,b), Chung et al. (2006), Egging and Gabriel (2006), Wu and Chen (2010), Egging (2013), Huppmann (2013), Baltensperger et al. (2016), Devine et al. (2016), and Lee (2016) |
| Capital investments in oil and gas projects and their valuation | Smith and Nau (1995), Smith and McCardle (1998), Clyman et al. (1999), Smith and McCardle (1999), Kenyon and Tompaidis (2001), Brandão et al. (2005), Hahn and Dyer (2008), Wang and Dyer (2012), and Gülp |
| Pipeline design | Brimberg et al. (2003), Andre et al. (2009), and Brimberg et al. (2007) |
| Valuation of pipeline capacity | Secomandi (2010b) and Secomandi and Wang (2012) |
| Pipeline capacity expansions | Andre et al. (2009), Garcia and Shen (2010), and Egging (2013) |
| Natural gas pipeline maintenance and repair management | Brito et al. (2010) and Angalakudati et al. (2014) |
| Optimal management of natural gas distribution through pipelines | Guldmann and Wang (1999), De Wolf and Smeers (2000), Rıos‐Mercado (2003), Dempe et al. (2005), Kalashnikov et al. (2010), Dempe et al. (2011), Schreider et al. (2014), Fodstad et al. (2016), and Hiller et al. (2018), Kirschstein (2018) |
| Open research questions | |
|---|---|
| Operational | Policy |
| How can natural gas pipeline capacity be better managed to serve the increasing demand from the electric power industry? | How should the contracts and market rules be redesigned to better facilitate the transactions between pipeline companies and gas‐fired power plants? |
| How should the flexibility in natural gas supply networks be valued? | Do existing markets provide sufficient price signals for necessary infrastructure development of pipelines and storage? |
| How should the increasing natural gas demand from the power sector be incorporated into expansion planning of natural gas transportation networks and storage assets? | How should the coordination between electricity and gas markets be improved (e.g., by aligning the market clearing times, coordinating scheduling mechanisms, etc.) to reduce the degree of financial risks experienced by gas‐fired generators? |
| How should the increasing volatility of demand be taken into account when managing natural gas storage operations? | |
| How should LNG storage be managed and valued? | |
Electric Power Industry Research Frontier and Potential Research Directions for OR/MS
The incredibly challenging task of perfectly balancing the supply and demand of electricity at all times opens up many research questions that can be addressed with the tools and techniques of OR/MS. Indeed, the electric power industry research has long been an important stream in OR/MS scholarship. The most well‐studied topics include capacity investments in generation assets, optimal scheduling and dispatching of power generation resources, electricity market design and its effect on economic and environmental outcomes, optimal bids by generation assets in forward markets, equilibrium behavior of electricity markets in terms of prices, production, emissions, and capacity investments, and optimal operation of generation assets. These topics are addressed with a wide range of methodologies including dynamic programming, stochastic programming, mixed‐integer programming, stochastic equilibrium models, dynamic games, robust optimization, and agent‐based simulations. For a review of common optimization methodologies used in energy research, see Banos et al. (2011).
Since the goal of this study is to highlight research opportunities related to the recent changes and challenges in the electric power industry, we do not provide a detailed review of all these actively studied topics. Instead, after providing basic definitions in section 3.1, we discuss several key areas that present the most relevant and current research problems for OR/MS scholars. Table 1 provides a sample of references for topics that are not thoroughly discussed in the review.
Definitions and Background Information
Historically, electricity has been generated and distributed in markets regulated by public utility commissions. In the regulated markets, vertically integrated utilities own the electricity supply chain from generation to the end‐customer. Because electricity rates are set by the public utility commissions, these markets provide price stability and long‐term certainty. However, consumer choice is restricted because utilities in regulated markets are natural monopolies. Since the 1990s, there has been a movement toward energy deregulation (or restructuring) around the world. Deregulation at the wholesale (i.e., generation) level means that independent power producers own generation assets and compete in the wholesale electricity market to sell electricity. Deregulation at the retail level means consumers can choose to purchase electricity from suppliers other than the incumbent utility. Although the rising prices and the California electricity crisis of early 2000s paused deregulation efforts in some states, more than two thirds of the electricity demand in North America is served by deregulated markets today. 1 Similar trends have been observed around the world including the majority of European countries, and many Asia‐Pacific countries such as Australia, New Zealand, Singapore, Japan, and South Korea.
The power grid is among the most complex technological systems built by humans. It serves to distribute electricity from generators to consumers with minimum energy loss while maintaining reliability and established service levels. This requires a hierarchical voltage structure, where power is generated with higher voltages, then distributed to consumers with lower voltages.
Power systems face the problem of determining a generation schedule to meet the load at the lowest cost possible subject to a large number of operational constraints. There are two related short‐term optimization models for this purpose.
ISO/RTOs commit the generators with the lowest price offers (bids) based on the needed demand for the next operating day to provide adequate notice of generation expectations. This results in 24‐hour clearing prices, which are also called the
ISO/RTOs are responsible from making sure that there are adequate generating resources to meet the daily demand curve plus a predetermined
In electricity markets, forward and future contracts are as essential as the immediate delivery through the spot market (i.e., the “real‐time market”). A forward contract is an over‐the‐counter agreement between two companies where one party is obligated to buy and the other to sell a specified quantity at a fixed price on a given date in the future. Future contracts are similar, but with the important distinction that they are traded on exchanges, making them standardized contracts. Forward and future contracts help market participants hedge their price risk. Market participants may also use these contracts for speculation; that is, they may increase their exposure to price risk in exchange for a risk premium.
There are several structural differences between the US and European electricity markets. Pool models in the United States are central dispatch models, which enable generating companies to offer price‐quantity pairs for the power supply in day‐ahead or intra‐day settings. In most of Europe, however,
Electricity Markets
The electric power industry is going through dramatic changes from the fall of natural gas fuel prices to the increasing deployment of renewable energy. There is one overarching research question related to all of these changes: How to redesign the grid and electricity markets to help society transition to a cleaner and more efficient electric power industry? It is widely agreed that the current electricity grid is inadequate to support such transition. Over the last decade, there has been a growing interest in “smartening the grid” by taking advantage of the improvements in advanced information and communication technologies. The smart grid vision entails making the grid more efficient and reliable through the use of sensors, smart meters, and other technologies that facilitate system monitoring to optimize efficiency and enable self‐repair (Koenigs et al. 2013). Smart grid also encompasses the development of new standards and market mechanisms that can support not only renewable integration but also the integration of other
The OR/MS field has a long history of building energy market models to provide decision support for policy makers (Hogan 2002). After the restructuring of the markets in 1990s, there has been significant interest in electricity market modeling. Earlier work in this area started off by building game theoretic models in order to investigate the strategic behavior in the newly restructured markets, including the formation of spot and forward prices, the equilibrium levels of supply from different generation technologies, capacity investments at equilibrium, the effect of market power on these equilibria and the potential for collusion and market manipulation (see the first row in Table 2). Since these models tend to be highly complex, a significant contribution by OR/MS scholars has been providing methodological advancements for the computation of equilibria (e.g., Anderson and Hu 2008, Gabriel et al. 2009, de Haro et al. 2007, Hobbs and Pang 2007, Hu and Ralph 2007, Shanbhag et al. 2011, Wilson 2008, Yao et al. 2008).
Electricity pricing has always been at the core of electricity market research (e.g., Banal‐Estañol and Micola 2009, Garcia et al. 2005, Joskow and Wolfram 2012, Lavigne et al. 2000, Peura and Bunn 2015, Wang et al. 2007, Yao et al. 2007). In deregulated markets, suppliers of electricity submit their price and quantity bids to a market operator who determines the optimal dispatch and the resulting payments to the suppliers. There is a strong body of research that studies the design and implementation of auctions in these markets looking at various aspects, such as bidding behavior, computation of market clearing prices, and the efficiency of the resulting dispatch (e.g., Derinkuyu 2015, Elmaghraby 2005, Madani and Van Vyve 2015, Meeus et al. 2009, Oren and Ross 2005, Pritchard and Philpott 2005, Reguant 2011, 2014, Schummer and Vohra 2003, Singer 2002, Toczy lowski and Zoltowska 2009).
One highly discussed issue after the transition to a competitive electricity market is under what conditions “energy only” payments are enough for generating firms to recover their fixed costs. Traditional marginal cost pricing may be problematic in the electric power industry, because total costs for power plants tend to be non‐convex due to factors such as start‐up costs, shut down costs, and minimum supply requirements, which make the marginal costs smaller than the average costs. As a result, some firms may fail to recover their costs through energy payments (i.e., payments for electricity delivered to the grid) only, which is known as the “missing money” problem (e.g., Cramton et al. 2013, Joskow 2006). Various alternatives to marginal cost pricing have been proposed in the OR/MS literature (see the discussion and references in Araoz and Jörnsten 2011, Bjørndal and Jornsten 2008, Liberopoulos and Andrianesis 2016 and O'Neill et al. 2005). Some of these alternatives are based on elevating the payments above the marginal cost, whereas others involve side payments.
The missing money problem is an active area of discussion in the electric power industry. There is an increasing concern that liberalized electricity markets do not provide sufficient incentives to build adequate generation capacity (Cramton et al. 2013). Wholesale electricity markets have price caps that constrain how much generating units can make during times of supply scarcity. As a result, the existing pricing schemes may not adequately reflect the value of investment in the necessary resources. Many argue that generating units need to receive compensation for their capacity in addition to the energy payments they receive by selling electricity (Cramton and Stoft 2005). These payments are handled in a capacity market. A capacity market supplements the revenues a generator can get from the energy and reserves markets, encouraging further capacity investments to ensure future resource adequacy. Several countries around the world (e.g., Colombia, UK, Germany) and several ISO/RTOs in the United States (e.g., PJM, MISO, NYISO) have been experimenting with capacity markets.
Large scale integration of wind and solar will likely exacerbate the missing money problem since these resources tend to depress spot electricity prices, making it even more difficult to adequately compensate generation resources. There has been much discussion related to market designs that spur adequate investment into conventional generators as well as renewables (Cramton et al. 2013, Cramton and Stoft 2005, Joskow 2006, 2008, Milligan et al. 2014, Pritchard et al. 2010). OR/MS scholars have also made contributions in this area. For example, Morales et al. (2014) and Pritchard et al. (2010) propose energy‐only market designs that account for intermittent generation by wind and solar plants. Ehrenmann and Smeers (2011) compare the investment behavior under energyonly and capacity‐only markets taking risk and risk aversion into account. Jiang et al. (2016) develop capacity metrics that can be used to quantify the capacity contribution of intermittent generators as well as storage resources by using an envelope‐based modeling method adapted from network calculus used in telecommunications research. Hach et al. (2016) develop dynamic capacity investment models to assess the effect of different capacity market design options including no capacity markets, capacity markets for only new capacity and capacity markets for both new and existing capacity. Their results show that, contrary to what critics of capacity markets suggest, a capacity market can actually lower the total bill of electricity generation for customers as it can reduce the potential to exercise market power. Yet, they also argue that capacity markets should only be introduced if the market is relatively close to a capacity shortage. The problems related to redesigning electricity markets to support renewable integration (or the integration of other distributed energy resources) is certainly not limited to the design of capacity or energy only markets. We refer readers to Hiroux and Saguan (2010), Milligan et al. (2014), Hu et al. (2018) for more detailed summaries of market design issues related to renewable integration, and to Ventosa et al. (2005) for a review of electricity market modeling trends. OR/MS scholars can contribute to this discussion of electricity market design by extending the existing methods in several directions, including more comprehensive treatment of market and technological uncertainties, and consideration of distributed energy resources.
One of the much‐discussed goals in the power industry is increasing demand response, that is incentivizing consumers to reduce their electricity usage during peak periods (for a literature review, see Siano 2014). The obvious and arguably most significant benefit from demand response is the reduced need for peak capacity. In addition, demand response can also be used as a solution to mitigate the effects of supply fluctuations, which can be a significant benefit in scenarios with high percentage of intermittent (variable) energy resources such as wind and solar. In one of the few OR/MS papers that incorporate demand response, Ströhle and Flath (2016) study this benefit and design a local market mechanism for matching flexible demand and uncertain supply. Valogianni and Ketter (2016) discuss the design of effective demand response programs based on a real‐world pilot program. Other OR/MS work in this area focus on the design and valuation of interruptible service contracts 4 (Baldick et al. 2006, Hatami et al. 2009, Oren 2001), design of energy buy‐back programs to reduce peak demand (Ding et al. 2012), and consumer level decision making (Zakeri et al. 2014).
Despite the well‐known benefits of demand response and despite the advancements in smart meters and incorporation of demand response programs in several utilities, the full potential of demand response has not been realized. One essential step to increase demand flexibility is to shift from flat rate electricity tariffs to tariffs that better reflect the varying cost of electricity, such as time‐of‐use pricing and real‐time pricing. Joskow and Wolfram (2012) discuss opportunities and challenges in introducing dynamic pricing. When achieved in large scale, a shift to dynamic pricing can have profound effects on long term capacity investments by utilities, technology mix of electricity generation portfolios, and environmental outcomes, such as emissions. There is a burgeoning interest from the OR/MS community to look at the effects of different electricity tariffs (e.g., Ata et al. 2016, Dong et al. 2017, Kök et al. 2016). The anticipated advancements in grid digitization will make it possible to adopt more granular pricing schemes (i.e., real‐time pricing) in the future. Ito (2014) shows that in a nonlinear pricing scheme, consumers tend to respond to average price rather than marginal or expected marginal price, likely because the cognitive effort required to understand complex pricing is prohibitively high. However, this response pattern may change with the introduction of smart home systems that track and display consumption and pricing information in real time and can take action on behalf of building owners according to their preferences. Together with the advancement of distributed energy resources, the increased demand elasticity may take us to a power system where demand follows supply rather than supply following demand, but such profound change can only happen with a complete makeover of the grid.
One step toward the advancement of distributed generation (such as rooftop solar PVs) has been net metering programs. Net metering (or net energy metering) is a practice that credits owners of distributed generation units for the electricity they add to the grid. Net metering can significantly reduce the payback period for acquiring a distributed generation unit (Gagnon et al. 2017); as a result it can encourage further market penetration of distributed energy resources. However, the net effect of these programs on distributed energy deployment is highly sensitive to the rate structure (Darghouth et al. 2016).
Net metering programs are welcomed by renewable generation advocates, but their adoption is not without controversy. Arguably, net metering at full retail electricity price allows distributed generation owners to avoid their full share of fixed utility infrastructure costs, which could trigger utilities to increase retail prices in order to recover these costs (Borlick and Wood 2014). Higher prices might further encourage distributed generation adoption, creating a feedback loop (Cai et al. 2013, Costello and Hemphill 2014, Darghouth et al. 2016). There is an ongoing discussion on the possibility of a “utility death spiral” with the idea that declining profits as a result of distributed energy resources would diminish the ability of electric utilities to survive (e.g., Castaneda et al. 2017, Costello and Hemphill 2014, Eid et al. 2014, Satchwell et al. 2015). Although theoretically possible, many argue that the utility death spiral is unlikely as utilities would find a way to work with regulators to ensure their viability (Costello and Hemphill 2014, Eid et al. 2014). Nonetheless, it is generally agreed that there are many economic and institutional implications of the move toward a decentralized power industry that need to be thoroughly discussed. Though there are important discussions on net metering in energy policy circles (Cai et al. 2013, Darghouth et al.2011, 2014, Eid et al. 2014), the topic is not extensively studied in the OR/MS literature. One exception is Hu et al. (2015), who incorporate the effect of net metering programs on the optimal renewable energy capacity investment at firm level. They show that optimal capacity investment indeed increases with the net metering rate, especially if the rate goes above 60% of the retail electricity price.
The electric power industry can learn from platform markets in the anticipated transformation process. A platform is essentially a business ecosystem that exploits two‐sided network effects (Parker and Van Alstyne 2005, Rochet and Tirole 2003) to match producers and consumers who wish to transact (Eisenmann et al. 2006, Parker et al. 2016). The grid of the future is likely to evolve as a platform that will bring together providers of distributed energy resources and consumers of electricity. A platform market can improve the integration of distributed energy resources into the planning and operation of electricity distribution systems through reduced transaction costs, improved provision of price forecasts, data analytics, and other smart technology services (Tabors et al. 2016). For example, a platform market would advance the adoption of electricity storage by facilitating more granular pricing that reflects the location‐ and time‐specific value of various services offered by storage. Such integration can potentially result in better system efficiencies, more affordable and reliable service, and a climate‐friendly energy system. It would also involve dramatic changes to the current business model of utilities (Burger and Luke 2016, Sioshansi 2014, The MIT Energy Initiative 2013), which opens up many research questions for OR/MS scholars (see Table 3 for a summary of open research questions).
Renewable Integration
Renewable energy production has increased significantly in the last decade, both in the United States and around the world. Wind and solar installed capacity has doubled roughly every three years on average over the past 30 years (Trancik 2015). Increasing use of intermittent renewable generation such as wind and solar brings a host of new challenges to grid operators and creates new avenues of research for OR/MS scholars (see Table 4 for a summary of existing research and open questions in this area).
The main challenge of increased renewable generation is the “intermittency” of wind and solar generation. Intermittent energy resource means that the energy resource is not continuously available to generate electricity. Wind and solar energy are intermittent because solar energy is not available at night, and wind energy is not available when the wind does not blow. Intermittent resources are considered to be
Another challenge related to renewable integration is transmission constraints. Wind and solar farms are typically built in remote locations far away from population centers. That means adding a significant amount of renewable energy to the power grid requires building massive transmission lines. Studying the impact of these transmission constraints on renewable integration is an important area of research (e.g., Fischlein et al. 2013, Papavasiliou and Oren 2013, Qi et al. 2015, Sioshansi and Short 2009). An ongoing discussion in this area is how transmission network congestion is managed in zonal versus nodal markets. As explained in section 3.1, European markets are zonal whereas US markets are nodal. Nodal markets are argued to be superior to zonal markets for congestion management and renewable balancing (Leuthold et al. 2008). Specifically, zonal pricing in congestion management does not fully take into account the physical characteristics of transmission. Moreover, zonal markets are likely to suffer high levels of unscheduled flows due to increasing penetration of renewables (Bjørndal et al. 2018). In contrast, nodal pricing (a.k.a. locational marginal pricing) arguably gives better price signals by reflecting the value of scarce transmission capacity and nodal pricing is being considered the European markets as well (Adamson and Parker 2013).
Since the integration of intermittent generation brings new challenges to the grid operators, it is necessary to develop
Capacity investments in renewable energy assets face a great deal of market, technological, and policy uncertainty. There is a strong body of research by OR/MS scholars on studying these investment decisions 6 both at the firm level and at the power systems level (e.g., Alizamir et al. 2016, Bonneuil and Boucekkine 2016, Boomsma et al. 2012, Bruno et al. 2016, Ehrenmann and Smeers 2011, Hu et al. 2015, Murphy and Smeers 2010, Parpas and Webster 2014, Pineda et al. 2016, Shittu et al. 2015, Siddiqui et al. 2016, Wang et al. 2016, Welling 2016). For example, Hu et al. (2015) study an organization's one‐time investment in renewable generation such as a bank branch investing in a solar rooftop system or a hotel investing in a solar thermal system. The renewable capacity can be coupled with a conventional technology to form a capacity portfolio that is used to meet the stochastic demand for electricity. They show that performing capacity investment using the average efficiency of a renewable technology may lead to significant overinvestment in renewable capacity; hence data granularity matters when it comes to evaluating renewable investments. Boomsma et al. (2012) use a real options approach to study investment timing and capacity choice for renewable generation assets. Ehrenmann and Smeers (2011) study capacity expansion in electricity markets with wind power penetration using stochastic equilibrium analysis. This line of research not only helps with firm‐level decision making on capacity investments, but also can inform policy makers about the effects of different subsidies and market structures on renewable capacity investment. For example, Hu et al. (2015) argue that providing the same federal subsidy across different geographical areas is inefficient, and that a lower subsidy rate is needed for areas with high yield–demand correlation. Ehrenmann and Smeers (2011) show the importance of taking risk aversion into account when evaluating different market policies. Specifically, they find that although energy‐only and capacity‐only markets (discussed in section 3.2) behave similarly when evaluated in a deterministic framework, risk aversion of investors increases the shortage of capacity in energy‐only markets with a low electricity price cap. Boomsma et al. (2012) compare different support schemes for renewable investment and argue that the extended discussions in Norway about choosing between a feed‐in tariff and green certificates may have caused a delay in renewable investments. Using a similar real options framework as Boomsma et al. (2012), Welling (2016) shows that uncertainty in electricity price and governmental support may have counterintuitive effects on renewable capacity investment; for example, higher uncertainty may trigger higher levels of capacity investment and a decrease of support over time can lead to higher cumulative investment in the short term. Capacity investment has been a central theme in OR/MS research on the electric power industry. An interesting future direction for this line of research is examining capacity investment decisions to create zero‐energy buildings and microgrids that rely on distributed energy resources (e.g., Marnay et al. 2008, Kamdem and Shittu 2017).
Large scale integration of intermittent resources will require rethinking the design of electricity markets. Even common metrics for evaluating the cost of generation technologies such as levelized cost per megawatt‐hour can be problematic when dealing with the intermittency of wind and solar energy (Joskow 2011). One of the issues with wind and solar integration has been the difficulty to value the capacity contribution of these resources due to the variability, inflexibility, and uncertainty associated with their capacity (NERC 2011). Different methods have been proposed and used for this purpose without a clear consensus (e.g., Jiang et al. 2016, Milligan and Porter 2008, NERC 2011). Failing to compensate generation assets for their capacity contribution may hinder the necessary investments into these assets in the long run. This problem is by no means unique to renewable generation. In fact, integration of wind and solar power exacerbates the problem since these resources tend to depress electricity spot prices. This makes it difficult to adequately compensate both conventional and renewable resources. Please refer to section 3.2 for a review of these market design issues.
Another interesting open question regarding renewable integration is whether the large scale integration and operation of renewable generation is more efficient in regulated or deregulated electricity markets. On the one hand, deregulation adds more flexibility into the system by creating markets for ancillary services, which can better facilitate the integration of intermittent renewable generation. Further, deregulation at the retail level enables consumers to choose their source of electricity, which may increase the penetration of renewables based on consumers’ choice. On the other hand, the presence of intermittent generation makes cost recovery of power plants more challenging in a deregulated market, whereas in regulated markets, this is not as problematic since the utility can ask its public utility commission to pass through the costs of a plant to its ratepayers. There is need for more research to understand whether the benefits from deregulation outweigh the added challenges when it comes to renewable integration. In particular, there are opportunities for empirical research and comparative studies to provide insights on how well renewable resources have been integrated under these different regimes.
The desire to move towards a cleaner energy sector prompts policy makers to use a variety of mechanisms to support renewable investment (e.g., renewable portfolio standards, tax credits, feed‐in tariffs 7 ). These policy instruments have a significant effect on renewable energy deployment (Fischlein et al. 2010, 2014). A growing subset of renewable capacity investment research specifically focuses on the effect of renewable policy on market outcomes. For example, Ritzenhofen et al. (2016) compare the impact of these different support schemes on electricity markets by developing a dynamic long‐term capacity investment model. Franco et al. (2015) build a system dynamics simulation model to analyze the long term effects of the policies proposed in British electricity market reform on electricity supply and environmental quality. Alizamir et al. (2016) build a dynamic optimization model to set and update feed‐in tariffs for renewable energy producers in order to accelerate the deployment of renewable energy. They show that the commonly used strategy of setting tariffs that maintain the same level of profitability across years is suboptimal, and that policy makers should set either an ascending or descending profitability index depending on the strength of technological learning and diffusion effects. Finally, Siddiqui et al. (2016) study optimal target levels for renewable portfolio standards (RPS) and how those targets change under different market structures. They show that when the non‐renewable sector has market power, the RPS target is lower yet the social welfare is higher compared to a market with perfect competition. For a more detailed review on carbon policy and its implications on the electric power industry, please refer to section 3.5.
Another research question that will be of interest with significant levels of intermittent generation is about energy bidding by intermittent resources. Making energy commitments would be challenging for wind and solar producers due to the intermittent nature of their generation. The necessity to make energy commitments would also increase the complexity of scheduling the power generation at a wind farm (Lima et al. 2015, Zhang et al. 2013). Optimal energy commitments in forward markets have been studied in the energy literature for different types of power plants (e.g., Anderson and Philpott 2002, Bitar et al. 2012, Botterud et al. 2012, Dent et al. 2011, Morales et al. 2010, Neame et al. 2003, Pritchard and Zakeri 2003). There is also a growing stream of research in studying joint optimization of the operations of an intermittent power plant with electricity storage (e.g., Brown et al. 2008, Brunetta and Tina 2007, Castronuovo and Lopes 2004, Garcia‐Gonzalez et al. 2008, Kim and Powell 2011, Korpaas et al. 2003, Mauch et al. 2012). Most of these papers either overlook supply uncertainty or account for uncertainty through building scenario trees and performing sensitivity analysis. The study by Kim and Powell (2011) is one of the few papers that develop valid admissible policies that take uncertainty in electricity price and wind generation into account. The authors study optimal commitments by a wind producer that operates an energy storage within a Markov Decision Process framework and derive an optimal commitment policy for the special case of uniformly distributed wind energy. Optimization of energy commitments by intermittent generation assets (with or without storage) is relatively an underdeveloped topic to which OR/MS scholars can contribute, particularly in improving the treatment of uncertainty in the existing methods.
Risk Management in the Electric Power Industry
Electric power industry has gone through a major transition in the last two decades with deregulation and market restructuring reshaping the industry in many countries across the world. One of the most important consequences of the resulting move toward a competitive electricity market is the emergence of price volatility. Because electricity has largely been non‐storable and is subject to high demand uncertainty, the volatility in electricity price can be quite extreme. For example, as quoted in Falbo et al. (2010), while volatility in “ordinary” financial series is about 10%–20% of average prices, this figure can reach to 300%–450% in some electricity prices. This extreme volatility exposes market participants to significant risk. As a result, following market restructuring and deregulation, a wide range of risk management practices have emerged in the electric power industry (Eydeland and Wolyniec 2003, Sioshansi 2002). Today, hedging strategies like forward and futures contracts are standard parts of doing business for market participants.
While many of the risk management tools developed in financial markets can easily be applied to the electricity markets, the unique characteristics of electricity requires developing some new tools and solution techniques (Deng et al. 2001). Specifically, since electricity has been and remains difficult to store, traditional no‐arbitrage methods of valuing commodity derivatives have not been used directly. Another challenge is the need to value a range of cross‐commodity transactions (e.g., between electricity and the fuel to generate electricity), such as spark and locational spreads (Deng et al. 2001). These challenges led the way to a significant body of research studying electricity derivatives, some of which are discussed in this section.
One of the most common ways to hedge against price risk in electricity markets is signing forward contracts before spot market trading occurs. Signing a forward contract eliminates the price uncertainty for both parties. There has been a lot of interest in the finance and economics literature to study electricity forward prices and how they relate to the spot price (e.g., Bessembinder and Lemmon 2002, Bunn 2004, Carmona and Coulon 2014, Fleten and Lemming 2003, Longstaff and Wang 2004, Lucia and Schwartz 2002). Some of the seminal papers include Bessembinder and Lemmon (2002), who build an equilibrium model for electricity forward markets, and Longstaff and Wang (2004) who conduct an empirical analysis of electricity forward prices and find significant risk premia in long‐term electricity markets.
While forward contracts protect against the fluctuating electricity price, devoting the whole production (or procurement) volume to forward contracts in generally deemed suboptimal compared to a more integrated risk management approach that mixes spot and forward sales (or procurement). Falbo et al. (2010) study this latter approach by looking at optimal spot and forward market sale combinations for an electricity producer. Dong and Liu (2007) study forward contracting in the presence of a spot market for a non‐storable commodity using a Nash bargaining framework. Sen et al. (2006) integrate the unit commitment model with financial decision making by incorporating forward contracts and spot market transactions into the scheduling model. There has been a growing interest in OM/Finance literature on procurement from commodity markets with several papers addressing optimal procurement from forward and spot markets (Devalkar et al. 2011, Goel and Gutierrez 2011, Goel and Tanrisever 2017, Inderfurth et al. 2013, Kleindorfer and Wu 2003, Pei et al. 2011, Popescu and Seshadri 2013, Secomandi and Kekre 2014, Seifert et al. 2004, Wu and Kleindorfer 2005). Haksöz and Seshadri (2007) provide a review of the early papers in this literature. Most of these papers study commodities other than electricity. Extending this body of research to specific issues concerning the electric power industry presents opportunities for OR/MS researchers.
In addition to physical trading through forward contracts, electricity market participants also engage in financial trading to hedge market risks using instruments like futures contracts, options, and swaps (for a review, see Deng and Oren 2006). Vehviläinen and Keppo (2003) study optimal portfolio selection for a mix of electricity derivatives and show that with proper modifications, some of the methods of financial markets are applicable to electricity markets. Deng et al. (2001) study the valuation of electricity derivatives and apply their valuation results to value generation and transmission assets.
The early literature on risk management in electricity markets typically focused on electricity producers for whom the main risk is wholesale price volatility. For retailers or load serving entities, though, there is another layer of uncertainty that stems from load fluctuations, usually referred to as volumetric risk or quantity risk. Boroumand and Zachmann (2012) argue that purely contractual portfolios consisting of forward and futures contracts are not efficient risk management devices for hedging the volumetric risk. Several papers in the literature address hedging the joint price and quantity risk faced by load serving entities or retailers (e.g., Boroumand et al. 2015, Oum and Oren 2009, 2010, Oum et al. 2006). There are also papers that look at risk management for retailers through non‐market mechanisms. For example, Aïd et al. (2011) study vertical integration as a risk management strategy and show that vertical integration has a superior efficiency over forward hedging when retailers are highly risk averse. Downward et al. (2016) also look at risk‐averse retailers and study their contracting choices as well as entry decisions, making vertical integration a possible endogenous outcome.
Though much of the risk management discussion in the electric power industry focuses on hedging market risks like price and load uncertainty, these are certainly not the only risk factors affecting the industry. The electric power industry is shaped by numerous other sources of uncertainty such as regulatory, capital, and operational risks. In fact, according to a global risk management study (Accenture 2013), the top two risk factors that are perceived to be on the rise by the market participants are regulatory and policy risks. Electricity markets are still evolving with uncertainty about what regulatory framework may emerge in the future. Further, a change in subsidies or tax regimes may completely shift the market towards different electricity sources (see section 3.5). These uncertainties play an especially significant role in technology and capacity investment decisions. Certain operational risks are also on the rise. Sources of operational risk include terrorist attacks (conventional, cyber, and electromagnetic pulse attacks), increasing number of natural disasters (partly due to global warming), solar flares and grid (generation and distribution) failures. The Camp Fire in California and the resulting bankruptcy of Pacific Gas and Electric Company (PG&E) 8 shows that the impact of operational risks can be very severe.
The ever‐changing landscape of the electric power industry creates the need for a more sophisticated and comprehensive risk management function. The OR/MS research on risk management in the electric power industry is relatively thin (see Table 5 for a summary of existing research and open questions). There is opportunity for OR/MS scholars to do high‐impact research by contributing to the development of an integrated risk management function in the electric power industry.
Climate Policy and Its Effect on the Electric Power Industry
Increasing concerns over climate change has led to regulations of greenhouse gas emissions across many countries around the world. There are two distinct policies aimed at emissions reduction: cap‐and‐trade and emissions tax. Cap‐and‐trade is a market‐based regulation that sets a limit (cap) on greenhouse gas emissions and then lets firms trade the emission permits that allow a certain amount of emission. Emissions tax, in contrast, is not market based. It is a policy based on imposing a fee on the emission (typically carbon) content of fuels and is arguably easier to implement. Though both policies put a price on emissions, their implications can be quite different, one of the main differences being the nature of uncertainty under the two policies. In cap‐and‐trade, emissions prices are determined through market forces, which introduces cost uncertainty. While an emissions tax provides more certainty about cost of compliance, its environmental impact is uncertain since there is no cap on emissions. There is much discussion about which policy is the best for reducing greenhouse gas emissions (e.g., Drake et al. 2016, Goulder and Schein 2013, He et al. 2012, Wittneben 2009). Ultimately the design of the policy plays a decisive role in the performance outcomes.
The OR/MS literature has produced a significant number of papers on climate policy design and implications (e.g., Carmona and Hinz 2011, Caro et al. 2013, Chen and Kettunen 2017, Chen et al. 2011, Drake 2018, Drake et al. 2016, İşlegen et al. 2016, İşlegen and Reichelstein 2011, Jaehn and Letmathe 2010, Kanudia and Loulou 1998, Nyiwul et al. 2015, Siddiqui et al. 2016, Subramanian et al. 2007, Sunar and Plambeck 2016, Zhao et al. 2010). Some of these papers uncover policy implications that are fairly complex or counterintuitive. For example, Chen and Kettunen (2017) show that unlike the popular argument, uncertainty in climate policy may induce more capacity investment in renewable technologies. Their results provide an important input into the debate on whether retaining the flexibility to update emission targets is beneficial despite its negative effect of causing policy uncertainty. Similarly, Drake et al. (2016) show that emissions price uncertainty under cap‐and‐trade policy results in greater expected profit than achieved under emissions tax policy with constant emissions price, which contradicts the conventional wisdom that higher uncertainty diminishes value. Finally, Aflaki and Netessine (2017) show that charging more for emissions may inadvertently discourage investment in renewables. Many of these papers provide regulators better insights about firm level response to various forms of climate policy as well as the interrelationships between different firm level compliance strategies. For example, Subramanian et al. (2007) focus on the trade‐offs between different compliance strategies of a profit maximizing firm operating under a cap‐and‐trade policy. The firm can invest in pollution abatement, 9 procure emission permits, or adjust the production output. The paper shows that, even with a simple model, the interplay between these strategies results in a non‐trivial equilibrium. Kroes et al. (2012) also analyze different levers of compliance and their effect on firm and environmental performance in an empirical study. In other empirical work, Jira and Toffel (2013) analyze under what conditions suppliers are more willing to share their emissions information with buyers in an attempt to reduce emissions throughout the supply chain, Jacobs (2014) shows that voluntary emissions reductions can create shareholder value, while Fabra and Reguant (2014) show that emission costs are almost completely passed through to electricity prices.
An important area of discussion in climate policy is
A productive area for OR/MS researchers studying climate policy has been the impact of climate policy on capacity, R&D investment, and technology choice decisions (Aflaki and Netessine 2017, Chen and Kettunen 2017, Drake et al. 2016, Fan et al. 2010, Kettunen and Bunn 2016, Kettunen et al. 2011). Aflaki and Netessine (2017) look at the effect of carbon taxes on renewable generation investments and show that the effectiveness of climate policy depends on the intermittency of the renewable technology. Since renewables require backup generation, which usually comes from more emission‐intensive technologies, an increase in the carbon price can backfire. The study argues that reducing the effect of renewable intermittency (e.g., through storage) would increase renewable capacity investment and improves the effectiveness of climate policy. Drake et al. (2016) investigate the impact of different climate policies on a firm's technology choice and capacity portfolios. Several papers in this stream focus on the effect of uncertainty in climate policy. For example, Fan et al. (2010) study investment decisions under regulatory uncertainty and show that a failure to consider risk aversion may bias policy models. Kettunen et al. (2011) study how climate policy uncertainty affects firm level investment decisions and market structure evolution. They incorporate real options and risk constraints in a multistage stochastic optimization model and show that using this more detailed financial analysis framework may create very different results from a conventional economic analysis. Similarly, Chen and Kettunen (2017) investigate the effect of uncertainty in emissions targets on firm profits, consumer surplus, and cost of compliance. Interestingly, they find that uncertainty may induce higher capacity investment and higher expected consumer surplus. Finally, Kettunen and Bunn (2016) also look at climate policy risk but in the context of resource dependency in capacity investments. They show that policy uncertainty and risk aversion may result in resource dependent capacity investments whereby the existing resources of a company impact the valuation of new resources. Ignoring this resource dependency may result in inefficient design of subsidies and incentives as well as underestimation of future market heterogeneity.
Closely related to the previously mentioned work on capacity investment, several papers address the uncertainty surrounding climate policy in the context of energy technology R&D investments (Baker and Shittu 2006, 2008, Baker and Solak 2011, 2014). For example, Baker and Shittu (2006) analyze the optimal R&D investment policy in response to an uncertain carbon tax. Baker and Shittu (2008) discuss the importance of incorporating both endogenous technical change and uncertainty for determining optimal technology R&D policy. Finally, Baker and Solak (2011, 2014) develop a modeling framework for energy technology R&D portfolio management under climate change uncertainty. An important factor that affects not only R&D investment and technology choice, but also the trajectory of technological change in electricity markets is the (uncertain) learning rates for renewable energy or other emissions abatement technologies. Cost reduction through learning‐by‐doing can be a significant determinant of technology adoption and investment since it has a direct impact on energy firms’ long‐run profitability (Shittu 2014). For example, wind and solar energy have experienced significant cost reductions over the last decade that made them more competitive and attractive from an investment standpoint (Lazard 2018). Learning‐by‐doing process presents as a feedback loop where investment into a technology invokes cost reduction benefits from learning, which in turn triggers further investment into the technology. However, if the learning rate is slow, this cascading effect may be insignificant. Thus, the uncertainty in learning rates is central to energy technology investment decisions. Several papers in the climate policy literature discuss the uncertainty in learning and R&D in the context of technological change (e.g., Baker and Shittu 2008, Bosetti and Tavoni 2009, Gritsevskyi and Naki′cenovi 2000, Shittu 2014). OR/MS researchers can build upon this stream of research by endogenizing learning‐by‐doing in other contexts (such as capacity investment and generation planning) and improving the stochastic modeling of uncertain innovation and technical change.
In addition to studies on capacity investment and technology choice, OR/MS researchers have also studied the effect of climate policy on production, procurement, supply chain design, and facility location decisions (Cachon 2014, Gong and Zhou 2013, İşlegen et al. 2016, Krass et al. 2013, Park et al. 2015). For example, Gong and Zhou (2013) study the optimal emissions trading, technology choice, and production plan for a manufacturing firm and characterize that the optimal emissions trading policy is a target interval policy while the production policy is of base‐stock type. Park et al. (2015) study the effect of carbon costs on social welfare by focusing on the last mile supply chain, namely retailers and consumers. They show that when the retailer's profitability level is low, imposing carbon costs can significantly increase social welfare. Finally, Krass et al. (2013) study a profit maximizing firm's decisions on emissions control technology, production quantity, and price in response to various forms of climate policy. They find that the firm's reaction to an increase in emissions tax is non‐monotonous in the sense that while an initial increase may prompt a switch to a greener technology; further increase may result in the opposite.
Operations management scholars and practitioners can play an important role in reducing emissions through revising operations and supply chain management practices (Plambeck 2012). Corbett and Klassen (2006) argue that environmental excellence can be a key to improving operational excellence and discuss the future of environmental research in operations management. Tang and Zhou (2012) also discuss how OR/MS research can help firms to strike a balance between profitability and sustainability. Though a significant body of OR/MS research in this area has already accumulated (see Table 6 for a summary of existing research and open questions), there is still need for further research, particularly on studying the impact of policy uncertainty on firms’ compliance behavior and the resulting environmental outcomes.
Electricity Storage
In the last decade, there has been a growing interest in energy storage, partly because of the increasing use of intermittent energy sources like wind and solar. Certain electricity storage technologies are finally on the verge of becoming cost competitive with their dominant conventional alternatives and industry participants expect costs to decrease significantly in the next five years (Lazard 2015).
Energy storage can provide many different types of service for ISO/RTOs, for utilities, and for end‐users (Eyer and Corey 2010, Sioshansi et al. 2012). Primarily, electricity storage helps with renewable integration, helps deferring new investment in generation, transmission, and distribution capacity, provides backup power, energy
In order to accurately value energy storage, one needs to optimize its operation based on the particular type of service(s) under study. Optimizing the operation of energy storage (such as finding the optimal policy for charging and discharging a battery) is a promising area of research for OR/MS scholars (e.g., Brown et al. 2008, Granado et al. 2007, Harsha and Dahleh 2015, Jiang and Powell 2015a, Paine et al. 2014, van de Ven et al. 2013, Xi et al. 2014, Zhou et al. 2015). This type of problem has similarities with classical inventory management problems, yet with distinct challenges primarily due to conversion losses in charging and discharging, dissipation losses, and ramp constraints. Many studies in this stream of research do not properly incorporate the stochastic nature of this problem (e.g., Brown et al. 2008, Granado et al. 2007), which provides opportunities for OR/MS researchers. Harsha and Dahleh (2015) analyze the optimal management and sizing of energy storage in the presence of intermittent generation using a Markov Decision Process. They consider renewable generators that directly face demand (such as those in game parks and industrial complexes) where energy storage is used to reduce the cost of electricity consumed. They derive a dual‐threshold type of optimal storage management policy. Billionnet et al. (2016) use robust optimization to study optimal sizing of a standalone energy system with batteries and intermittent generation. Xi et al. (2014) develop a stochastic dynamic program to co‐optimize multiple uses of a distributed energy storage (e.g., a battery that has been installed in a home). Jiang and Powell (2015a) develop an approximate dynamic programming algorithm that can be applied to study the joint operation of intermittent generation and electricity storage. Zhou et al. (2015) study a merchant using storage to manage electricity surpluses in an electricity wholesale market and show how negative prices affect the optimal storage policy structure. Paine et al. (2014) develop a dynamic programming model of pumped hydroelectric storage operation under different market rules and show how market rules affect operational strategies as well as the profitability of the pumped hydro facility (We provide a more detailed review on hydropower operations in section 3.7). Finally, Jiang and Powell (2015b) study optimal bidding in the electricity markets by grid‐level electricity storage.
In addition to developing operational policies for electricity storage, OR/MS scholars have also focused on the use of storage for renewable integration. For example, Qi et al. (2015) adopt the perspective of a central planner and study the optimal location and sizing of electricity storage systems within a framework of transmission network planning. They show that even a small size electricity storage system may significantly reduce the necessary transmission investment for renewable integration. In a firm level study, Kim and Powell (2011) focus on renewable energy commitments in forward markets that have associated penalties if the contracts are breached. In this setting, storage is an insurance that minimizes the risk of making advance commitments in the market. On a more strategic level, Anderson and Parker (2013a) study co‐specialization investment decisions of an emergent storage technology to better complement the operations of a renewable technology. In their setting, storage smooths the volatility of power delivered by the renewable technology it complements.
For OR/MS scholars, there are many potential avenues of future research regarding electricity storage (see Table 7 for a summary of open questions). As mentioned above, one of the most fundamental unsolved problems is the valuation of storage based on multiple types of services. A consistent methodology capable of comparing the net value of multiple services provided by electricity storage to other incumbent technologies would have a significant impact on the adoption of these technologies. On the policy side, there are regulatory changes that must be adapted to enable widespread integration of energy storage (Sioshansi et al. 2012). Testing the effect of different market designs and regulations on the integration of energy storage is another promising area for future research. For example, it is unclear whether integration of storage is more efficient in a regulated or a deregulated market. Currently, it is hard for energy storage to compete in energy markets where the rules do not adequately value the flexibility that storage can provide, whereas it may be easier for a regulated utility to make the economic case for energy storage. However, as the platformization of the grid progresses in deregulated markets, the situation may easily reverse. On a final note, studies looking at the effects of storage deployment on emissions (e.g., Denholm et al. 2005, Greenblatt et al. 2007, Sioshansi 2011a) or on social welfare (e.g., Schill and Kempert 2011, Sioshansi 2010, 2011b, 2014) are mostly based on energy arbitrage use of storage. As markets evolve to accommodate multiple uses of storage, new studies will be needed to reassess the impact of storage deployment on emissions and social welfare.
Hydropower Operations
Hydropower is one of the oldest methods of producing power. It is low cost, low emissions, and flexible. All of these factors made hydropower the most widely used renewable source of energy around the world, accounting for over 16% of the world's net electricity production. 10 Pumped storage hydropower is also the dominant form of energy storage on the grid (see section 3.6 for a discussion of energy storage).
Hydropower is often seen as the perfect complement to wind power, as it can quickly ramp up and down and therefore can balance the fluctuations in wind output. With the growing percentage of wind energy in the power system, this complementarity has received a lot of attention. Many countries are investigating the opportunity to integrate wind and hydropower systems in order to optimize their output through coordinated operation (Acker 2011). However, coordinated operation of wind and hydro resources is not a trivial problem due to many non‐power constraints on hydro units. For example, hydro units may be required or prevented from operating at certain times to avoid flooding, to provide irrigation water, to maintain reservoir levels for recreation and for other ecosystem considerations such as fish passage. All of these considerations make up a challenging optimization problem, which provides important research opportunities for OR/MS scholars. There is already a burgeoning stream of research on wind and hydro integration (Belanger and Gagnon 2002, Benitez et al. 2008, Bueno and Carta 2006, Castronuovo and Lopes 2004, Jaramillo et al. 2004), yet with limited presence in OR/MS journals (e.g., Löhndorf et al. 2013, Steffen and Weber 2016, Vespucci et al. 2012). Wind‐hydro coordination problem is not just a technical matter that needs to be resolved within the domain of utilities. Since hydro units have many different functions (power generation, irrigation, recreation, etc.), there are typically numerous stakeholders involved with different market and economic considerations. The interactions between the many functions of hydro and the involvement of these different stakeholders will ultimately shape the prospects of wind and hydro integration. Thus, in addition to operational‐level studies addressing technical aspects of the coordination, there is need for more macro‐level studies, such as those based on dynamic simulation methods, to understand the future role of hydro in this context.
Though hydropower's potential benefits for wind integration has increased its value in the eyes of the utilities, its overall popularity may very well decline in the near future due to reduced availability of good spots for dam construction, concerns about the environmental impact of building new dams on river ecosystems, and more recently, concerns about the effects of climate change. A changing climate implies changes in evaporation rates, precipitation patterns, glacial melt rate, and the frequency of extreme meteorological events. All of these factors decrease water resources and hydropower potential in some regions while increasing them in others. Some negative effects have already been observed in various locations. For example, on the Colorado River, an increased rate of evaporation has resulted in reservoirs that are less than half full (Lustgarten 2016), in California a draught‐induced shift from hydropower to natural gas increased ratepayers’ spending by an estimated $1.4 billion between 2011 and 2014 (Gleick 2016), and in Brazil draughts caused rolling blackouts nationwide (Holthaus 2015). Though the changing climate is also expected to benefit hydropower in some regions (Acker 2013), the fundamental problem remains: Hydropower operations will face more uncertainty in the future and are going to become more challenging as a result of climate change. OR/MS literature already offers a strong body of research on hydropower optimization (see Table 8 for references and open research questions). Nonetheless, there is a need for revising these models and methods to better take into account the effect of climate change and the uncertainty that it brings.
With the changes in climate, the value of hydro becomes more uncertain as well, increasing the riskiness of hydropower capital investments. This is especially true for larger dams. There is some work in the OR/MS literature that focuses on capital investments in hydroelectricity, typically using a real options framework (e.g., Bøckman et al. 2008, Bruno et al. 2016, Thompson et al. 2004); yet, more work is needed to fully understand the effect of climate change on the desirability of these investments. Despite the potential concerns, many countries are still interested in expanding their hydropower resources in order to quickly meet low‐carbon energy generation targets (Holthaus 2015); but a higher reliance on hydro may actually increase carbon emissions in prolonged draughts as the gap in generation is likely to be picked up by potentially inefficient fossil fuel‐based generators. Thus, calculating the effect of hydropower capacity expansions on achieving climate targets is not straightforward and remains an open research question.
Electrification of the Transportation Sector
With increasing concerns about emissions, alternative fuel vehicles such as electric and hybrid electric vehicles have significant government support since these vehicles can use cleaner sources of fuel compared to conventional gasoline or diesel. Several studies have confirmed that plug‐in hybrid electric vehicles emit less CO2 over their entire fuel cycle than conventional vehicles (Parks et al. 2007, Peterson et al. 2011, Samaras and Meisterling 2008, Sioshansi 2012, Sioshansi and Denholm 2009, Sioshansi et al. 2010, Stephan and Sullivan 2008), although their net effect on other pollutants such as SO2 and NOX depends on the electricity generation mix and the charging pattern (Sioshansi 2012, Sioshansi et al. 2010). More importantly, grid‐connected transportation solutions like plug‐in electric vehicles will grow cleaner over time as they will benefit from any future reductions in generation emissions in the electric power industry.
Plug‐in electric or hybrid vehicles (PEVs) have a battery that can be recharged from the grid with a plug‐in charger. Thus, wide‐spread adoption of these vehicles could represent a significant potential shift in the use of electricity and the operation of electric power systems. In particular, the need for additional generation, transmission, and distribution capacity can increase, especially if vehicles are charged during periods of peak demand (Parks et al. 2007). There are many studies that examine the impacts of integrating PEVs into the power system (e.g., Lemoine et al. 2008, Parks et al. 2007, Sioshansi 2012, Sioshansi and Denholm 2009, Sioshansi and Denholm 2010, Sioshansi et al. 2010, Stephan and Sullivan 2008, Wang et al. 2010). Richardson (2013) provides a review of early work in this area. A key result from these studies is that the net effect of PEVs on power system costs and emissions largely depends on the charging pattern of the PEVs and the electricity generation mix. For example, if PEVs are charged during peak demand hours, they tend to use costly, less efficient, and dirtier generators than if charging is delayed to off‐peak hours. Several studies (e.g., Samaras and Meisterling 2008, Sioshansi and Denholm 2009, Stephan and Sullivan 2008) have demonstrated that if PEVs are charged during the night using excess generating capacity, they will have a minimal impact on the power grid in terms of capacity requirements. Thus, it will be important to incentivize desirable charging patterns through a market‐based approach by providing different tariffs, such as real‐time pricing and time‐of‐use rates. Several studies study the cost and emissions impact of PEVs under these different schemes (e.g., Sioshansi 2012, Wang et al. 2010). As PEVs become more wide‐spread, designing market mechanisms or regulations to invoke desirable charging patters will become an important area of research.
Another key aspect that shapes the effect of PEVs on the power system is the opportunity for vehicle‐to‐grid (V2G) services. With the technological advancements reshaping the grid, PEVs will potentially be able to provide many services to the grid including capacity and ancillary services. For example, PEV batteries are essentially electricity storage devices, which can be charged when the cost of generating electricity is low and discharged to the grid when it is high (Peterson et al. 2010). This capability improves the efficiency of the electric system by decreasing the peak demand, which reduces the usage of high cost, high emissions peaking generators. Further, with V2G capability, PEVs can provide other ancillary services such as regulation and spinning reserve. Using PEVs for these services would allow the system to operate more efficiently, decreasing the emissions from generation units that are currently used to provide such services (Kempton and Tomic 2005, Sioshansi and Denholm 2010). The potential impact of V2G services on the electric power industry is a newly developing area that presents interesting research opportunities. In one of the few OR/MS papers on this topic, Broneske and Wozabal (2017) study contracts between PEV owners and entities called
Despite governmental subsidies that support their adoption in several countries, electric vehicles (plug‐in or otherwise) have not achieved mass adoption yet, largely due to three limitations. The first one is the high upfront costs of these vehicles. Electric vehicles are costlier than conventional vehicles, mainly because of the expensive battery they contain. Even though the fuel costs are lower for these vehicles, the price difference is high enough to become a deterrent. The second reason behind lack of adoption is “range anxiety.” Electric vehicles have a shorter range compared to conventional ones before the need for refueling arises. Furthermore, it takes a couple of minutes to refuel conventional vehicles, while the battery of an electric vehicle can take several hours to recharge (Level 2 chargers), although level 3 charges such as the Tesla supercharger (135KW) can provide an 80% charge in 30 minutes (Srdic and Lukic 2019). Hybrid and plug‐in hybrid vehicles remedy this issue of range anxiety, as the vehicle can switch to gasoline once the battery is depleted. Finally, the third limitation is the lack of a recharging infrastructure, which presents itself as a chicken‐and‐egg problem: Drivers are reluctant to purchase electric vehicles until the recharging infrastructure is wide‐spread enough, whereas service providers are hesitant to invest heavily on infrastructure unless there is significant demand. In the United States, Tesla is the first company to substantially invest in a rapid (level 3) recharging network and, in order to capture the benefit of the investment, the system is available only for Tesla vehicles (Nicholas and Hall 2018). This chicken‐and‐egg problem presents interesting policy questions that can be analyzed through mathematical modeling: Should governments subsidize consumers to promote early adoption of electric vehicles (as many do), or the private sector to encourage investment in recharging infrastructure? How does the volatility in fossil fuel prices affect the adoption behavior and hence the effectiveness of any of these subsidies? OR/MS scholars can contribute to this literature by building sophisticated mathematical models of adoption dynamics for electric vehicles. See Table 9 for references to the existing research and a summary of open questions.
There is already a growing stream of OR/MS research studying mass adoption of alternative fuel vehicles. Struben and Sterman (2008) build a dynamic simulation model of the diffusion of alternative‐fuel vehicles by incorporating key feedback structures such as R&D, learning by doing, technological spillovers, and the development of refueling infrastructure. Lim et al. (2015) study the impact of range anxiety and resale anxiety on the mass adoption of electric vehicles using a stylized durable goods model. They compare different business models (such as battery owning versus leasing) in terms of electric vehicle adoption, emissions, consumer surplus and firm profitability; and determine which model works the best depending on the degree of resale anxiety. Gnann and Plotz (2015) and Gnann et al. (2018) provide reviews of models for market diffusion of alternative fuel vehicles and their refueling infrastructure. He et al. (2017) study PEV sharing platforms as an alternative to ownership, which potentially can remedy some of the aforementioned hurdles in front of PEV adoption. One important component that tends to be missing in this stream of research is the incorporation of uncertainty into the modeling process. Many papers study the effect of demand or technological uncertainty through sensitivity analysis rather than building a stochastic model. One exception is He et al. (2017) who use distributionally robust optimization to address the uncertainty in customer adoption of PEV sharing. Another notable exception is Cohen et al. (2015) who study the effect of demand uncertainty on optimal consumer subsidies for electric vehicles (and other green technologies). The authors analyze a variation of the price‐setting newsvendor model where the electric vehicle manufacturer sets the price and capacity investment while the policy maker sets the consumer subsidy levels. They show that policy makers can significantly miss their desired adoption targets if they ignore demand uncertainty. This finding confirms that there is indeed a need for models that incorporate uncertainty into the decision‐making process. This is an area where OR/MS scholars can contribute to with their expertise in building and analyzing stochastic models.
Another potential area of research for OR/MS scholars is the design and operation of the refueling infrastructure for electric and plug‐in electric vehicles. One solution proposed as an efficient refueling infrastructure is battery switching (also known as battery swapping) stations. In these stations, the drivers would be able to quickly exchange a depleted battery for a fully charged one, which would solve the issue of range anxiety as long as there is a dense network of these stations. Further, under such a system, the drivers would effectively lease the costly battery rather than owning it and would pay for usage based on miles driven. There are significant operational challenges in designing and running the network of battery switching stations. In terms of infrastructure design, the main challenge is uncertainty about adoption rates, which not only deters investment but also makes it difficult to plan for the spare battery requirements in these stations. Avci et al. (2015) build a model of switching‐station‐based electric vehicle system and compare it to conventional electric vehicles in terms of consumer adoption and environmental impacts. From an operations management point of view, they find the optimal inventory level for spare batteries and the optimal price to charge per mile driven. From a policy point of view, they show that policies that are more effective in reducing oil dependence (e.g., electric vehicle subsidies) tend to be less effective in reducing emissions. The reason is that such policies not only increase adoption but also lead to more driving of electric vehicles; and in most countries electricity generation is still obtained using carbon emitting technologies. Mak et al. (2013) study the planning process for deploying battery switching infrastructure using distributionally robust optimization. They build two models with different objectives and study the potential impacts of technology advancements and the standardization of the battery technology on optimal deployment strategy. Their paper is one of the few in this literature that takes uncertainty into account.
Other types of refueling infrastructure alternatives such as public charging stations have also been studied in the literature. He et al. (2013) study the optimal allocation of public charging stations for plug‐in electric vehicles by a central planner with the goal of maximizing social welfare. Their game‐theoretical model investigates the interactions between electricity price, availability of charging stations, and the destination choices for these vehicles and calculates traffic and power flow distributions. Many other papers discuss the optimal locations for refueling stations using various methods such as flow‐refueling location models (e.g., Capar et al. 2015, Chung and Kwon 2015, Lim and Kuby 2010, Yildiz et al. 2016) and agent‐based simulations Zhao and Ma (2016). Other work on PEV refueling infrastructure include Nurre et al. (2014), who study the optimal operations of battery switching stations with a model that considers V2G capabilities, Hung and Michailidis (2015), who use a queuing model to study the optimal routing of vehicles that request charging to stations with available charging resources, and other similar papers that adapt the vehicle routing problem to incorporate the new challenges introduced by PEVs (e.g., Doppstadt et al. 2016, Goeke and Scheider 2015, Hiermann et al. 2016). There is still a great deal of uncertainty about the kind of refueling infrastructure that will take‐off once alternative fuel vehicles become more wide‐spread, and despite the initial excitement they generated, battery switching stations may never become the dominant model. One potential area of research for OR/MS scholars is comparing different types of refueling infrastructure in terms of their effectiveness of promoting the adoption of PEVs and studying the effect of different market and policy scenarios on their success.
Natural Gas Industry and Its Impact on Electric Power Generation
The OR/MS field has a long history of research in the context of natural gas industry. Table 10 summarizes some of the major research topics in this area that have been actively pursued by OR/MS scholars. Natural gas plays an increasingly important role in power generation. In our review, we highlight research areas that have become more relevant in light of this increased dependence between the natural gas and electric power industries noting that the overall opportunities of research regarding natural gas industry are not limited to what is covered in this section.
The recent increase in the supply of natural gas because of higher production from unconventional resources has significantly reduced natural gas prices in the United States, and thus, increased the appeal of natural gas‐based electricity generation. Further, with increased penetration of intermittent renewable generation, the responsiveness of gas‐fired plants has become a critical asset in balancing out the variability in wind and solar power. Indeed, currently, more than 30% of electricity generation in the United States uses natural gas as fuel (EIA 2016), which has made the electric power industry the second largest consumer of natural gas. 11 This shift has profound implications for both the electric power and the natural gas industry. One of the pressing issues is the inadequacy of transport networks for natural gas. In the United States, the transport capacity (e.g., pipeline capacity) for natural gas has not kept up with the increase in supply and demand. For example, the Algonquin Gas Transmission pipeline that traverses New England, New York, and New Jersey has run at 100% capacity for more than four years due to increased reliance on natural gas for electricity in that region (Patel 2016). In areas that heavily rely on natural gas for electricity, this inadequacy of pipeline capacity can quickly become problematic since generators can be left with no means to receive gas on days when there is high demand. Further, scarcity of transport capacity may contribute to price surges and volatility for natural gas. For example, in late January and early February 2013, spot prices at the Algonquin and New York gates were nearly an order of magnitude higher than the Henry Hub reference prices. 12 Such high natural gas price volatility is challenging for utilities to manage; thus, greater coordination between electric power and natural gas industries is needed to mitigate the effects.
The growing interdependence between electric power and natural gas industries, especially in the United States, has drawn a lot of attention from energy industry practitioners and scholars. Hibbard and Schatzki (2012) discuss the many challenges associated with the growing interdependence and provide a framework for electric–gas coordination. The facts that electricity market must be balanced in real time and that the two markets are not designed to clear simultaneously make coordination between the two markets particularly challenging. Academic research that studies electric–gas coordination is growing, albeit with limited contribution from OR/MS journals. Many of the papers in this literature focus on optimizing the coordinated operation of electricity and natural gas networks (e.g., Alabdulwahab et al. 2015, Bai et al. 2016, Chiang and Zavala 2016, Cui et al. 2016, Li et al. 2008, Liu et al. 2009). For example, Chiang and Zavala (2016) show that coordinated dispatch of natural gas and electricity transmission systems can result in significant improvements in economic performance and flexibility. Alabdulwahab et al. (2015) look at the coordination of electricity and natural gas infrastructures for the purpose of firming the variability of wind energy using a stochastic unit commitment model. Bai et al. (2016) also take into account wind uncertainty and use an interval optimization‐based model to optimally coordinate the operations of the integrated natural gas and electricity systems. Another stream in the literature focuses on coordinated expansion planning of electricity and natural gas infrastructures (Chaudry et al. 2008,2014, Unsihuay‐Vila et al. 2010, Zhang et al. 2015). Overall, the literature on the coordination of natural gas and electric power industries is relatively small, presenting many opportunities for further research.
Another consequence of the shift to natural gas‐based generation is the significant changes in the demand pattern for natural gas. Specifically, the variability and uncertainty in intermittent generation is likely to increase the variability and uncertainty in natural gas demand. The overall cyclicality of electricity demand is also bound to reshape the demand curve for natural gas. All this variability requires flexibility in natural gas supply networks, as well as better management of pipelines and storage. This opens up a lot of relevant problems for OR/MS scholars. Below, we briefly review the existing research on the management of natural gas distribution and storage, and discuss several open questions in light of the increasing interdependence of the natural gas and electric power industries.
Natural Gas Distribution
Pipelines are the major delivery mode for natural gas and play a critical role in matching the supply and demand of natural gas across different locations (Secomandi 2010a). In the United States, pipeline capacity has not kept up with the recent increase in natural gas supply and demand, which can cause serious problems for utilities in areas that highly depend on natural gas for power generation. Reliability becomes especially a concern on the coldest days during which pipeline constraints force home heating to compete with power generation for scarce gas supply. For example, in New England, natural gas is the dominant fuel for generating electricity, but the natural gas pipeline system within the region is relatively small. In unusually cold winters such as the winter of 2017–18, the region experiences large spikes in natural gas price and wholesale electricity price largely due to pipeline capacity constraints. 13 Pipeline constraints adversely affect not only the power system reliability, but also the emission levels since during instances of insufficient natural gas supply or spikes in natural gas price, the grid operators tend to switch to higher‐emitting options such as power plants using oil or coal. Given these consequences of pipeline capacity becoming a binding constraint, research on the management of pipelines becomes ever more relevant.
Pipelines are operated by pipeline companies, who provide transportation services to shippers of natural gas (producers, merchants, or local distribution companies). Pricing (or valuation) of these services is an important problem in practice. Secomandi (2010a,b) analyzes the valuation of pipeline capacity for different players in the industry using a real options framework. He focuses on point‐to‐point contracts and shows that pipeline capacity should be priced at its trading value. Secomandi and Wang (2012) study the valuation of network contracts that allow merchants to ship natural gas among more than two locations. Their method outperforms those in practice because it is based on using an optimal operating policy for the underlying contract. Possible extensions to this stream of research include incorporating market power or transaction costs into the analysis, and adapting the existing framework for transactions with gas‐fired power plants.
Another stream of research related to managing pipeline capacity deals with capacity expansion investments (e.g., Andre et al. 2009, Egging 2013, Garcia and Shen 2010). There are strong parallels between capacity investments into electricity transmission networks and investments into natural gas pipelines; thus, power systems literature on transmission network expansions is highly relevant to this stream of research. For example, Garcia and Shen (2010) use an equilibrium model to study capacity expansion of critical energy infrastructure, including gas transportation networks, though their main application area is electricity transmission networks. Coordinated expansion planning of electricity and natural gas infrastructures (e.g., Chaudry et al. 2008,2014, Unsihuay‐Vila et al. 2010, Zhang et al. 2015) is an important area of future research. For more potential areas of research, we refer the reader to survey papers on optimization models for natural gas distribution (Ríos‐Mercado and Borraz‐Sánchez 2015, Zheng et al. 2010).
Market rules and contract designs that govern the transactions between generators and pipeline companies can play a big role in ensuring power system reliability, especially in areas with high natural gas demand and inadequate pipeline capacity. Current market rules may not be sufficient to efficiently reconcile natural gas supply and demand (Peress and Karas 2017). Consumption profile of gas‐fired generators is not consistent with historical pipeline operational design. Specifically, most transportation services are based on the principle of uniform hourly flow whereas very few gas‐fired plants need steady flows of gas. While ways to overcome this fundamental inconsistency have been devised (such as multiple nomination periods in a day to allow gas‐fired plants to change their delivery schedules), it is doubtful that under the current market rules the existing capacity and flexibility is optimally used. Further, there is no standardized market construct for pricing various forms of delivery flexibility exercised in practice. As a result, it is argued that markets do not receive the correct price signal to channel investment into the much needed delivery flexibility (Hibbard and Schatzki 2012). On a related note, many gas‐fired generators contract for interruptible pipeline services, which may hinder their ability to produce electricity when natural gas demand is high. Yet, firm transportation contracts that improve reliability, which are costlier, are not incentivized by most electric wholesale power markets (INGAA 2014). In areas where adequate natural gas supply is crucial to power system reliability, adjustments to market design may be necessary to enable gas‐fired generators to recover the cost of firm pipeline capacity. In short, there are many market design and contract optimization issues to be resolved to facilitate the transactions between generators and pipeline companies, which present research opportunities for OR/MS scholars.
Natural Gas Storage
There is a significant body of research in OR/MS journals that deals with optimal valuation and operation of natural gas storage. Natural gas storage plays a key role in matching supply and demand of natural gas. Demand for natural gas fluctuates significantly; yet, the production of natural gas cannot be immediately adjusted to meet the fluctuations in demand, making storage a critical component of the supply chain. The primary use of storage is mitigating the seasonality of demand by storing excess gas produced in summer to meet the peak demand in winter. Natural gas storage is gaining more importance due to the aforementioned changes in the demand structure, which brings about new opportunities for research to OR/MS scholars.
In a nutshell, managing a natural gas storage requires determining an inventory trading policy that tells the merchant how much to buy from the wholesale market and inject into the storage facility, or withdraw from the storage and sell into the market, based on the current natural gas spot price and inventory levels (Secomandi 2010a,b). Among the first in the literature to analyze this problem, Secomandi (2010a,b) finds that the optimal policy is characterized by two time and spot price dependent base‐stock targets. Lai et al. (2010) extend this work by considering a multifactor forward curve model for price instead of the one‐factor mean reverting spot price model.
The problem of evaluating the real option to store natural gas has been studied in the finance, energy, and operations management literatures extensively. Analytical valuation of storage options typically does not exist because of injection and withdrawal constraints. Various computational methods have been developed for storage valuation, including numerical partial differential equation techniques (Chen and Forsyth 2007, Thompson et al. 2009), binomial and multinomial trees (Felix and Weber 2012), approximate dynamic programming methods based on Monte Carlo simulation (Boogert and De Jong 2008, Carmona and Ludkovski 2010, Lai et al. 2010), and approximate linear programming (Nadarajah et al. 2015). Practitioners typically employ two heuristic policies to value seasonal energy storage; the rolling‐intrinsic (RI) approach and the rolling basket of spread options approach (Eydeland and Wolyniec 2003, Gray and Khandelwal 2004). Lai et al. (2010) find that both heuristics have near‐optimal performance. Wu et al. (2012) identify the conditions under which the RI heuristic deviates from optimality and develop methods to bring the RI heuristic closer to optimality. Secomandi et al. (2015) use the RI heuristic to study the effect of futures term‐structure model error on the valuation and hedging of natural gas storage and to propose approaches to remedy the negative effects of model error on hedging. Finally, Lai et al. (2011) is among the few papers that study the valuation of LNG storage at a regasification terminal. The main difference between natural gas storage and LNG storage lies in the fact that the inflow of commodity into the storage facility is controllable in the natural gas case, but not in the LNG case. With the increasing importance of LNG in the natural gas market, there is going to be a need for further research in valuation and operations of LNG terminals and storage.
Conclusion
In this study, we have surveyed over 500 articles published in the OR/MS literature with the goal of identifying those papers that address particular problems regarding the electric power industry and with the goal of promoting research into a number of critical areas that can benefit from the expertise of the OR/MS community. Topics surveyed include renewable integration, energy storage, electricity market design, risk management in the electric power industry, the effect of climate policy on the electric power industry, hydropower operations, the electrification of the transportation sector, and the links between electric power and natural gas sectors. These areas are all changing rapidly because of the increased competitiveness of renewable energy technologies, constantly evolving climate policy, improved hydrocarbon production technology, the lag of gas transport infrastructure buildout, and the increasing digitization of the electric grid with the implication that power demand might soon become much more price elastic. We identified opportunities for contribution to both theory and practice in these areas and highlighted more than fifty open questions. Some of the overarching open questions that will shape the future of the electric power industry are also summarized below in Table 11.
Electric Power Industry Overarching Open Questions
| Electric power industry overarching open questions |
|---|
| How should market prices reflect the varying cost of electricity? What is the potential impact of dynamic pricing on capacity investments, demand response adoption, emission levels, and technology mix of electricity generation portfolios? |
| Can digital platforms be developed to improve the integration of distributed energy resources into electricity distribution systems? If platforms are widely adopted, how will incumbent utilities’ roles change and what business models might allow them to remain viable? |
| How can the electric power industry develop an integrated and comprehensive risk management function that encompasses not only the market risks but also policy risks? How can the industry better integrate financial and operational hedging for improved risk management? |
| How will improved understanding of climate change affect the operation of existing assets as well as future capacity investments? What policies are likely to be robust to alternate climate scenarios and how can policy be better coordinated with operational decision making? |
| How will electricity storage impact transmission and generation investment, power markets, and emissions? Should markets be redesigned to enable widespread integration of electricity storage? What is the interplay between technology change and policy? |
| What policies might foster coordinated hydropower asset investment and operations with renewables to help balance variable generation? How should policy makers reconcile the needs of the diverse hydropower stakeholder body? How does climate change uncertainty affect hydropower operations, capacity investments, and the role hydropower can play in achieving climate targets? |
| What will be the effect of the transition to electric vehicles on the grid and on achieving climate targets? Should policy makers allow/encourage manufacturers to sponsor proprietary recharging networks or will joint network sponsorship better solve chicken and egg adoption issues? |
| Given increasing natural gas demand from the power sector, how should the coordination between electricity and gas markets be improved? What policies should be adopted to increase transportation capacity and storage assets? What are some key possible unintended consequences of policy changes? |
During this revolutionary time in the electric power industry, we believe that the OR/MS community is in a strong position to provide valuable decision‐making support by bringing the necessary operational considerations into the discussion of electric power policy matters. OR/MS scholars are especially well‐equipped to solve problems at the intersection of business, economics, and policy. Although there are large literatures devoted to the electric power industry that focus on technical issues and literatures that focus on policy matters, we believe that OR/MS scholars can play an important role by bridging these focal areas and by recognizing the mutually interacting and dual‐causality dynamics between operations and public policy.
Footnotes
Acknowledgments
We thank the many seminar participants who provided feedback, especially Ekundayo Shittu, Elizabeth Wilson, and the Industry Studies Association conference attendees. In addition, we are very grateful to the editors and the review team whose careful reading and guidance throughout the review process greatly improved the paper.
1
The 2018 State of the Utility, Annual Survey Report by Utility Dive.
2
Alberta (AESO), California (CAISO), Texas (ERCOT), Ontario (IESO), New England (ISO‐NE), Midcontinent (MISO), New York (NYISO), Pennsylvania New Jersey Maryland (PJM), and Southwest Power Pool (SPP) are the ISO/RTOs in the United States and Canada. The rest of Canada and Southeast and Northwest regions of the United States are traditionally regulated markets.
3
Power exchanges in Europe include Netherlands (APX), Belgium (Belpex), Germany (EEX), Austria (EXAA), Scandinavia (NordPool), Poland (PolPX), and France (Powernext).
4
5
6
7
A
8
PG&E and its parent company were sued in November 2018, for not properly maintaining its infrastructure and equipment, which led to the Camp Fire, the deadliest wildfire in California history. The lawsuit claims that a failure in a transmission line was the cause of the fire. On January 29, 2019, PG&E filed for bankruptcy in response to the financial challenges, which is estimated to be more than $16.5 billion, associated with the fire. The chief executive of the company resigned in the same month.
9
Pollution or emissions abatement refers to any measure taken to reduce pollution or emissions.
10
11
12
13