Energy Storage for Mitigating the Variability of Renewable Electricity Sources

Marc Beaudin; Hamidreza Zareipour; Anthony Schellenberg; William Rosehart    Department of Electrical and Computer Engineering, University of Calgary, 2500 University Drive NW, Calgary, Alberta, Canada T2N 1N4

Acknowledgments

This work was partially supported by NSERC Canada and the Institute for Sustainable Energy, Environment and Economy (ISEEE).

1 Introduction

This chapter provides a survey of applying electric energy storage (EES) for facilitating the large-scale integration of variable renewable electricity sources (VRES), such as wind and solar power, into electric power systems. Large-scale integration of VRES introduces significant uncertainty into operation and planning of electric power systems. Electric energy storage is considered a tool for mitigating the impacts of VRES uncertainty [1] and [2].

In general, several criteria are analyzed when considering and choosing EES technologies for a specific application [3–5]. Those criteria include lifetime, life cycle, power and energy, self-discharge rates, environmental impact, cycle efficiency, capital cost, storage duration, and technical maturity. Based on these criteria, the appropriateness of EES for various applications has been evaluated in the literature, such as, for flexible alternating current transmission systems, small-medium-large-scale applications, system efficiency, emissions control, peak shaving, and deferring facility investments in peaking generators [3–8]. A limited amount of the reported research focuses on the necessary characteristics of EES specifically for VRES applications [4–6]. Nevertheless, other energy-storage literature can be applied for this purpose [8–11]. However, these references can either have a limited scope of EES or be outdated [4,9], and [10]. For example, [4,6], and [10] do not elaborate on battery types, and [4,8,11] are limited to bulk EES only.

In the present chapter, the previous literature is extended by providing an updated review of the state of technology and installations of a broad range of EES technologies. The main focus is on applications and appropriateness of each EES technology for mitigating the variability of renewable electricity generation sources. This chapter also includes discussions on important criteria for understanding the potential future of EES technologies, considering the impact of distributed generation, maturity and timing, and world metal resources on EES penetration for VRES applications.

The remainder of this chapter is organized as follows: Section 2 provides an overview of variable renewable electricity sources. Section 3 presents the current state of EES systems, which includes information on current worldwide installations, cost of the technology, and current applications. Section 4 discusses the appropriateness and competitiveness of EES for VRES applications. Finally, Section 5 provides a summary of the chapter.

2 An Overview of Variable Renewable Electricity Sources

With growing concerns about the environmental impacts of the electricity sector, there has been increasing interest to invest in wind and solar power. The 121 GW of global wind-installed capacity in 2008 produced 260 TWh of electricity and saved 158 million tons of CO2 [12]. It is estimated that the worldwide wind cumulative capacity reached 318 GW in 2013 [12]. In the same year of 2008, nearly 6 GW of new photovoltaic and thermal solar to power installations contributed to the cumulative installed capacity of 14.7 GW [13]. In recent years, installed solar to power capacity has been growing very fast (e.g., 8% growth in 1992 and 46% in 2008), and reached 139 GW in 2013. However, the variable nature of these renewable resources introduces a new source of uncertainty in the operation and planning of electric power systems.

Variations in VRES depend on the size of the evaluated system and the timescale of wind variations. Proportionately, small wind farms tend to have larger expected hourly variation than variations from an entire area. For example, in Western Denmark, it can be reasonably expected that wind power may vary by 3% of its 2400 MW capacity, whereas a 5 MW wind farm in the same area may vary by 12% [14]. VRES timescale variations can be characterized as microscale, mesoscale, and macroscale. Microscale variations primarily affect regulation (seconds to minutes), while mesoscale variations affect the load-following timescale (minutes to hours), and macroscale variations affect the unit-commitment timescale (hours to days). While microscale fluctuations are smoothed to a significant extent across a typical wind-power array, mesoscale and macroscale fluctuations can be significant for wind farms and even for an entire region [15]. For example, in Denmark, small hourly variations of 10% are common for wind farms, but hourly fluctuations above 30% still occur once every 1.1 years (1/10000 probability) [15]. As a comparison, in the Republic of Ireland, the probability of a thermal-generation plant tripping in any given hour is between 0.0006 for the most reliable units to 0.003 for the least reliable units (median time to failure of 9.6-48 days) [16].

The impact of large-scale VRES variations on power systems differs by time-scale [17,18]. In microscale, large-scale integration of VRES may require significantly more regulation reserves and frequency control depending on the power systems characteristics. For example, a study conducted for Ontario Power Authority indicates that integration of 10,000 MW wind-power capacity into the Ontario system of 26,000 MW peak demand would require an 11% increase in regulation requirements [19]. In mesoscale, VRES variations impact the balance between the supply and demand, and thus, may require a significantly increased amount of operating reserves [16]. The same study for Ontario shows that a 47% increase in operating reserves is necessary in order to deal with mesoscale variations of wind under a 10,000 MW wind-integration scenario [19]. In macroscale, VRES variations impact unit commitment and scheduling of conventional generators, and unpredictable variations may result in significant economic costs. For example, system start-up costs could increase by up to 227.2% in German power systems as a result of day-ahead wind-power forecasting errors [20]. In addition, the inverter-based operation of wind generators also has some power-quality impacts on power systems [21], which are further discussed in Section 4.1.1.

In practice, large variations of VRES, particularly wind, have led to operational difficulties in some cases. As an example, on February 26, 2008, an unexpected 1,400 MW drop in wind-power generation coincided with an unexpected load increase and loss of a conventional generator in Texas [22]. These events forced the Electric Reliability Council of Texas (ERCOT) to take emergency steps and cut 1,100 MW firm load in order to restore system frequency. In addition, wind generators were dispatched down three times in 2008 in the Irish power system for security reasons [23]. Such events and considerations are the basis for limiting large-scale wind-power integration in some power systems. For instance, a 362 MW wind-power restriction is in effect on the 800 MW peak load power system in the Canary Islands [24]. In the province of Alberta, Canada, the electric system operator put a 900 MW cap on wind-power integration because of “operational concerns” [25]. Although this cap was later removed, the system operator has been continuously looking for solutions to deal with the variability associated with wind power in Alberta, such as significant investments in a central wind forecasting project.

Some utility studies have concluded that the grid can absorb wind energy up to 10% of the system load without major technical changes or significant costs [14] and [16]. However, the same studies have also recognized...