by Dan Sowder, Duke Energy
The interconnection of distributed solar photovoltaic (PV) generation to the utility distribution system (12-24 kV) is an increasing trend across many utility systems. For a distribution grid designed for one-way energy flow, this can pose operational and reliability challenges. To make these challenges more complex, it is envisioned that solar PV is likely to proliferate in the form of many small residential rooftop installations interconnected on a distribution feeder. In contrast with a large solar PV installation at a single point of interconnection on a feeder, hundreds of smaller solar PV installations will be connected at various points. This additional level of dispersion of solar can compound some of the challenges of absorbing distributed solar on the grid.
Energy storage shows promise in addressing some issues. As utilities find themselves at the confluence of improving energy storage technology and a growing awareness of distribution circuit problems caused by solar, insight into how energy storage can be deployed and operated on the distribution grid is needed.
Duke Energy’s Emerging Technology Office has identified energy storage as an important emerging technology and has deployed several diverse systems to develop understanding of this technology and its benefits. Based on excerpts from a technical report, this describes one of Duke Energy’s energy storage field tests and details some initial results on the effectiveness of the system and its current application. This project was made possible by the strong collaboration between Duke Energy, FIAMM Energy Storage Solutions and S&C Electric Company.
Description and Overview
The Rankin Energy Storage System (ESS) was conceived to understand how energy storage can address distribution circuit problems caused by distributed solar PV. The project was designed to provide insight into what value streams can come from a distribution-deployed energy storage system, the costs of deployed energy storage, and how well this system fits into utility operations.
The Rankin ESS was installed within the fence line of Duke Energy’s Rankin Avenue Retail Substation in Mount Holly, N.C., in late 2011 and was commissioned in early 2012. The Rankin 1208 circuit was chosen because it has a 1.2-MW solar PV facility some 3 miles from the substation. This rooftop solar facility is owned by Duke Energy and is interconnected onto the 12.47-kV circuit. Large, rapid fluctuations of the solar PV output have been observed at the solar interconnection point and in the substation load profile. During periods of low circuit demand and high solar output, a significant portion of the circuit load is being supplied by the solar output, resulting in reverse power flows on the distribution line.
The Rankin ESS was designed to serve as a proof of concept for technology development and demonstration. As a result, many components of this system were selected because of their developmental potential, broad applicability to energy storage system and cost considerations.
The battery system is composed of 12 FIAMM SoNick sodium-nickel Zebra batteries connected in parallel in a dedicated 20-foot container as shown in Figure 1.
The total system nameplate capacity is 216 kW/282 kWh. These batteries originally were designed for use in electric and hybrid electric vehicles and have been in service in this application for more than 120 million vehicles miles. Sodium-nickel batteries are attractive for the power swing mitigation application because of their higher power-to-energy ratio because larger power output is needed for a short time.
Power Conditioning System (PCS)
The primary functionality of the PCS is to accomplish bidirection ac and dc conversion, ensure high-quality ac and dc power is delivered, and to manage the system communications and controls. The PCS hardware is composed of an S&C Electric Company PureWave Storage Management System (SMS) with a capacity of 1.0 MW/1.25 MVA.
Interconnection Balance of System
The interconnection balance of system includes hardware necessary to interconnect the other components to one another and to safely interconnect with the distribution circuit including the interconnection transformer, overhead switches to provide a means of visible disconnection from the grid, and auxiliary power.
A closed-loop control algorithm drives the system using telemetry provided by existing substation relays and other components. No direct telemetry is provided by the solar facility; only local circuit parameters are needed. This removes the cost and data latency associated with collecting telemetry from multiple, potentially distant, distributed solar facilities.
The algorithm first collects the circuit’s real-power load from the substation relay and identifies when there is a large rate of change of this load. A smooth version of circuit loading is calculated and the necessary battery command is issued to achieve this smoothed circuit loading. This process is repeated on a continuously cycling loop that repeats every 500 milliseconds–1 second.
Locating the energy storage system at the substation rather than with any individual solar installation enables mitigate load fluctuations caused by distributed generators on the circuit, regardless of their level of dispersion along the circuit. Also, the centralized location enables the storage system to fit within the existing utility footprint of a substation, take advantage of existing substation equipment and easily demonstrate other applications that can be more difficult with distributed batteries. The primary shortcoming of this location is that it prevents the ESS from effectively decreasing reverse power flows along the circuit–and thus preserving a typical voltage profile–because it cannot absorb energy until it reaches the substation.
System Performance Analysis
To evaluate the effectiveness of the energy storage system while in operation on the grid, performance data from nine days during late summer and fall 2012 was analyzed. These nine days were selected because of the large solar output fluctuations experienced.
The circuit load data was divided up into 1-minute segments. Within each segment, the highest and lowest power points were measured. The difference between these two kilowatt levels represents the 1-minute power fluctuation. This parameter, Swing kW, was calculated for each of the 6,479 1-minute segments within the nine days that were analyzed. This calculation was repeated using 1,295 5-minute time segments from the same nine days.
This calculation was performed on the measured, smooth circuit load curve and the calculated, unsmooth circuit load curve that would have occurred if the battery were not installed. The difference between Swing kW with and without the battery gives an indication of how effectively the energy storage system reduced power swings in the circuit load.
Figure 2 graphs the power swings that would have been experienced without the battery vs. the mitigated power swings that were measured with the battery online. Points that fall below the center line indicate the battery reduced the magnitude of power swings within its time interval. Linear regressions are shown for both the 1-minute and 5-minute data.
Assessment of System Effectiveness
Two major conclusions can be drawn. The energy storage system successfully reduced the magnitude of solar-induced power swings in most cases. More specifically, 81.10 percent of the 1-minute Swing-kW data points were below the slope line (5,253 of 6,479 points) and 90.89 percent of the 5-minute Swing-kW data points were below the slope line (1,177 of 1,295 data points). Not all Swing-kW data points fall below the slope line, indicating that sometimes the energy storage system makes the circuit power swings larger; that is undergoing further analysis. This generally occurs in response to relatively small solar-induced power swings, typically below 400 kW. This analysis reveals that the Rankin ESS successfully reduced the magnitude of a majority of solar-induced power swings but occasionally increased the magnitude for relatively small solar-induced power swings.
Optimal Energy Storage Capacity Assessment
A review of the effectiveness of the battery in mitigating load fluctuations vs. the size of the original power swing shows that the system consistently reduces power swings measured up to around 500 kW before the system starts becoming less effective because of capacity limitations.
Figure 3 plots the average percentage reduction in power swings vs. the size of the original power swing. The graph shows the system becoming increasingly more effective for power swings up to about 400 kW, at which point it becomes less effective. This implies that an energy storage system of a given capacity (in this test, typically 200 kW) can effectively reduce solar-induced power fluctuations consistently at a capacity of roughly double the energy storage capacity (in this test, between 400-500 kW).
This analysis has shown that a substation-based energy storage system can reduce the magnitude of solar-induced power fluctuations from the perspective of upstream power system components in most cases. It also suggests that an energy storage system can, on average, consistently mitigate power fluctuations above power capacity.
These findings will be used to further develop the system’s control algorithm to include other benefits such as dispatchable reactive power. Also, these results will be used to evaluate how this system reduces the cost of integrating distributed solar to the grid via enhanced substation transformer life span, improved circuit voltage stability because of more effective load tap changer operations, and the avoidance of additional protective relaying equipment.
The Rankin ESS continues to be used for testing new algorithms and applications including enhanced power fluctuation mitigation algorithms, voltage control algorithms, and other applications such as frequency regulation. The continued advancement and demonstration of energy storage on the grid is critical for technology to be adopted by utilities at a large scale.
An Integrated Energy Storage Checklist
by Jennifer A. Eirich, EnerSys
Utilities are turning increasingly to batteries for grid stabilization and integration of renewable energy sources. Battery storage systems are a convenient, flexible and cost-effective solution that store energy in off-peak times and allow utilities to quickly ramp up and provide surplus energy during peak demand. Installing an effective integrated system involves a series of important decisions.
The Five C’s
Chemistry: various forms of lead acid and nickel cadmium to high performance thin-plate pure lead and lithium ion. The key to making the right choice is understanding application requirements. Key questions include:
- How much space is available? Is power density a concern?
- What is the required life cycle? How accessible is the battery?
- What is the physical environment? Will the battery be subjected to extreme heat or cold that might lead to degradation?
- What is the anticipated frequency and depth of discharge?
In some applications, long life cycle is critical; in others, it’s costly overkill. Lithium-ion batteries can offer more than twice the life cycle of a lead acid battery. At almost five times the cost, however, it is only cost-efficient for those applications in which size, weight and longevity make it necessary. Often, lead acid batteries deliver the necessary performance requirements at greater savings than other high-energy storage solutions.
Conditioning: Choosing the appropriate ac system is a science and requires at least as much due diligence as specifying the appropriate chemistry. Beware of battery agnostics–vendors or integrators who specify the battery system and give the impression that power conditioning is a one-size-fits-all solution. Ideally, it’s best to seek a dc supplier, such as EnerSys, that offers a range of storage chemistries and has experience partnering with ac experts. This should provide a range of choices based on the merits of the individual technologies rather than force fitting a single solution.
Control: Storing energy involves understanding what to do when there is failure, how to control it and how to avoid potential catastrophic results. The more energy stored in a device, the more will be released in a catastrophic failure. Electrical shorts also generate heat that can spread to surrounding cells. Some systems monitor individual cells; others monitor strings of devices. The former is more effective at spotting spikes, identifying trends and avoiding thermal runaway. In addition to the amount of isolation, other variables include switching/response time, power accuracy, dynamic transfer, configuration and scalability.
Cover: Don’t compromise when specifying the enclosure that will protect your investment. Start with a robust structure–a watertight and sealed enclosure on a thick concrete pad with a grounding structure. Dc ground fault monitoring and protection provides safety, security and reliability. Consider other built-in safety features, such as fire suppression, spill containment and gas detection, as well as a climate control system to optimize battery life.
Cash: When seeking financing, be sure to account for the project needs. Think beyond installation to include training and the first year of maintenance.
Dan Sowder is senior project manager in Duke Energy’s Emerging Technology Office and has been with the company since July 2010. He holds a bachelor’s of science degree in aerospace engineering from the U.S. Naval Academy. He has a MBA from the University of North Carolina and a master’s degree in engineering management from Old Dominion University. Sowder is a registered professional engineer in electrical power systems engineering. Reach him at 704-832-8642 or firstname.lastname@example.org.