Thursday, March 11, 2010

SDM Best Thesis Prize awarded for grid-scale energy storage research - SDM Pulse Spring 2010

By John Kluza, SDM ’08

Editor’s note: John Kluza was awarded the SDM Best Thesis Prize in October 2009 for his thesis, "Status of Grid-Scale Energy Storage and Strategies for Accelerating Cost-Effective Deployment."

The electric grid is so ubiquitous in the modern world that its presence and functionality are taken for granted. However, there are increasing challenges to the continued success of the electric power system, including the growing need for dependable electricity, the desire for improved system efficiency, the influx of intermittent renewable generation, and the limitations of aging, expensive grid infrastructure.
As a student in MIT’s System Design and Management Program, I became keenly aware of these issues during my summer 2008 internship at A123 Systems, a lithium ion battery manufacturing startup that was founded on MIT technology. While there I learned that work was under way to build and pilot grid-scale energy storage systems using A123’s batteries. Large-scale energy storage promises to solve many of the grid’s current problems, so this project—based on one of a variety of emerging storage technologies—fascinated me and got me curious about how such systems might be deployed cost-effectively.

This started me down the path of my thesis topic. First, I identified all the unique benefits that could be produced by grid storage, based on a variety of secondary sources and discussions with forward-thinking utility representatives. I also gathered information on what financial benefits could be produced. These benefits could be found throughout the electric grid value chain— from generation to transmission and distribution (T&D) to customer loads (see Figure 1).
Most of the energy storage applications fell into two categories: 1)"energy-oriented" applications that could be accomplished using long-discharge batteries (for example, storing low-cost energy produced off-peak and delivering it back to the grid during peak periods) or 2)"power-oriented" applications that require fastresponding, brief-discharge batteries (for example, frequency regulation, which resolves momentary imbalances between electric generation and load).

Each of these categories has its own system requirements, and each individual application also has further unique requirements of the grid storage hardware. Through my secondary research, I developed a list of 14 energy-storage applications with estimated financial benefits ranging from $72/kW installed over 10 years to $1,649/kW installed over 10 years.

I then identified which types of energy storage technology could be used for these applications and, for the purposes of my thesis, constrained my work to distributed systems with more than five minutes of discharge time (denoted by the red circle in Figure 2). The major types of technologies investigated were sodium sulfur batteries, flow batteries of the zinc-bromine and vanadium redox type, lithium ion batteries, advanced lead acid batteries, and high-speed flywheels.

Through discussions with storage manufacturers as well as secondary research, I identified the unique advantages and estimated cost of each technology. The capital costs for these systems varied widely, from $370/kW to $4,000/kW measured by power, or from $347/kWh to $8,000/kWh measured by energy. Systems were generally more appropriate either for power or energy applications as reflected by the capital expenditure (capex) per unit of energy stored or power capacity. Systems were also evaluated using another metric that is more meaningful for comparing energy applications to natural gas peaker plants and natural gas combined cycle gas turbine (CCGT) plants. It is called cost per kWh cycled, and it measures both efficiency and cycle life as well as capex per unit of energy stored (Figure 3).
Figure 3: Capital expense per kWh cycled

Finally, I distilled the majority of actual or proposed distributed grid-scale energy storage options into six classes, estimating the maximum benefit produced for each based on the combination of applications that could be supplied simultaneously. I then reviewed each class from a technical feasibility and regulatory perspective and estimated the cost of a grid storage system based on each type of technology. The expected maximum benefit was compared to the expected cost to identify which combinations of applications and technologies could potentially produce a cost-effective installation. Due to the approximate nature of the available data points, the estimation was restricted to: likely to lose money (-), likely to roughly break even (0), or likely to offer a positive net present value (+), as shown in Figure 4 for the installation classes and technologies investigated.
Figure 4: Profitability expectations

This research led me to draw a number of useful conclusions, including identifying many markets for grid storage of varying maturity, size, and value. Frequently the best approach to deploying a cost-effective installation in these markets will include combining applications that have easy-to accrue benefits. Additionally, while not necessarily grid scale, displacing oil- or diesel-fired generation is often cost-effective and can be an entry point for suppliers.

More specifically, I found that there are classes of installations that may make economic sense given the proper conditions. For example, the power-oriented market is attractive now with existing technology. The most accessible application in this market is ancillary services, such as frequency regulation, because energy storage is exceptionally well suited to provide it and there is an open, cash market for these services in some regions of the United States. Lithium ion systems can currently be used cost-effectively for ancillary services and in the future can potentially provide community energy storage. High-speed flywheels are expected to become attractive for ancillary services in a few years. (See Figure 4.)

I found that the energy-oriented market opportunity is currently limited both for cost and regulatory reasons. Distributed energy storage systems still need to reduce their cost per kWh cycled to be competitive. Also, for many applications there is no clear mechanism for paying the owner of the grid storage system.

Nevertheless, sodium-sulfur batteries, currently the most common technology, are attractive in the near term for industrial energy management in many regions, including the United States. Other applications are also attractive in foreign countries, such as Japan, that have different electric system constraints and regulations, including renewable power management, wholesale load shifting and T&D capacity deferral. Additionally, zinc-bromine flow batteries are expected to be attractive in the near- to midterm for T&D capacity deferral, industrial energy management, and renewables management.

Across the board, it is complicated for the owner, such as a utility, to accrue all the financial benefits generated by the grid storage system due to the interdependent nature of the electric grid. Some benefits are produced as avoided costs instead of cash, and the regulations governing the electric grid are not always conducive to creating and collecting the benefits. The importance of regulation and government policy in making these systems economical cannot be overstated.

Though there are many challenges for storage on the grid, I am optimistic that energy storage will ultimately strengthen the grid and enable cleaner, less expensive electricity. I hope that this research will help to clarify the topic for readers so that more work can be done on the topic, accelerating the deployment of grid-scale energy storage.

No comments:

Post a Comment