In Part 1, we discussed the usefulness of batteries in managing the grid while mentioning that battery performance can be hard to quantify when placed behind the utility meter. In Part 2, we will look at how battery charging strategies must be planned with sufficient foresight of system needs.

While contemplating new technology investments, it is often worthwhile to consider if there are any unintended consequences. On the small scale it could look like this: If I buy a new iPhone, do I receive a more expensive phone bill due to increased usage? For battery investments, the unintended consequences have more to do with strategically controlling the charging and discharging.

For the purposes of this article, the difference between direct and indirect effects center on the time in which they occur. Direct effects occur in conjunction with an event dispatch instruction (see Figure 1). An indirect effect occurs at some other point throughout the day (see Figure 2). These effects need to be understood and accounted for by the grid operator in order to provide adequate forecasts and properly identify the appropriate resource mix to balance generation and demand.

Essentially, the storage system should make the customer load more predictable and not deviate from the normal schedule until receiving the dispatch instruction. The customer load should then return to the normal schedule after termination of the DR event so as to minimize indirect and unintended impacts that may necessitate more issues and reactionary events. As far as I know, the naming conventions (e.g. direct rebound, indirect, etc.) used within this piece are original.

Direct Rebound Following Performance

Even successful direct rebound (DR) events often see a slight direct rebound effect, a demand profile greater than average immediately following the event. This slight increase in usage immediately following an event may be attributable to an air conditioning compressor working hard to return a large facility to a comfortable temperature after responding to a DR event by shutting off for an hour or more. Maybe the usage increase is the result of a battery charging back up after a response to a DR signal caused a discharge below the necessary capacity to maintain functionality for other services, such as peak demand shaving. Regardless of the reason, this type of rebound is shown in Figure 1.

Figure 1: Direct Rebound Effect Example
 

This type of direct rebound should be accounted for when modeling DR events. Often, any negative effects of a quick direct rebound are far outweighed by the benefits associated with the performance of the event. Still, depending on the application and the sensitivity of the local electric infrastructure, direct rebound effects may stress the margins and nullify the benefits of the DR event.

Indirect Performance Effects

During the recent Battery DR Pilot project, there were days when the storage system was called upon to perform, with the initial assumption that the batteries would perform in the same fashion each day and could, therefore, be forecast into the system. On days when the utility signaled the storage device to perform (event days), the storage device often deviated from the range of how it performed on previous days even during hours that it was not expected to perform. It is possible that the typically programmed schedule, which is optimized to minimize energy and demand charges via peak shaving techniques, was manually overridden on test event days prior to the expected performance period. Unfortunately, this deviation from the standard schedule in the timeframe before and after test events had a significant impact on the projected load profiles that the DR event was intended to improve. Event day performance of a battery storage system can change the value proposition and necessity of the event, to begin with.

For the pilot project, the schedule of events was pre-established weeks in advance. Similarly, day-ahead energy market awards would have known dispatch and restoration times. Building a resource stack to balance electricity supply and demand is predicated on known usage profiles that assist decision makers in finding the most cost-effective answer. Increasingly, utility decision makers and end-use customers are turning to energy storage as a cost-effective tool in their resource stack. However, the performance of the storage in the hours outside of the focus timeframe must also be considered so as to avoid creating different, possibly more extreme, system balancing issues.

Figure 2: Example of Indirect Effects of Battery Energy Storage
 

Does the behavior of the storage device in order to achieve the desired performance outweigh the benefits associated with that desired performance? In Figure 2, does the charging that occurs from 8:45-10:00 disrupt the predictive models in such a way that it changes the necessity of the DR performance? Does the system now have enough demand to justify utilization of a different resource? What are the effects of that unexpected demand after the event? These are questions that must be considered in today’s electricity architecture. In other words, we must ensure that the battery is charging and discharging to maximize its potential. This isn’t a new idea, but the application is important. When asking a storage device to perform differently than typical, we must consider how it achieves that performance, not just when or if it will perform.

These indirect effects can be controlled, modeled and maximized. Energy storage can be a great asset to improve grid efficiency, but only if deployed with a comprehensive outlook.

In addition to the effects from charging and discharging laid out above, the behavior of energy storage resources also impacts the grid in the following meaningful ways.

Indirect Rebound Effect

The idea of indirect energy impacts was articulated in the June 2016 United States Data Center Energy Usage Report from Lawrence Berkeley National Laboratory. Within that report, the authors lay out that ICT (information and communication technology) direct energy consumption, the energy consumed by the ICT devices and infrastructure themselves, “is likely the simplest and ultimately the least important ICT energy effect [although it made up 1.8% of total U.S. electricity consumption]… the indirect energy effects are likely to be of much greater magnitude, owing to the breadth of the various mechanisms by which ICT services alter energy use.” To more fully understand that claim, it is important to appreciate how indirect energy effects become complex and far-reaching very quickly. The authors mention how teleconferencing, enabled by ICT advances, has replaced some of the need to fly across the globe for meetings. With respect to energy storage, the effect is not as straightforward. Energy storage is part of the decentralization paradigm shift in electricity generation enabling customers to own their own renewable energy generation. To an extent, energy storage can displace conventional fast response generation resources that adjust output to balance electricity supply and demand, but there are many more factors.

Another example outlined in the US Data Center Energy Usage Report described how a less costly e-book might lead consumers to purchasing more books. Consumers could also apply these savings to other goods and services, whose prices themselves are dependent on consumer income. These indirect rebound effects consider “consumer and market responses to the energy efficiency improvement. Such consumer and market-wide responses are likely to occur because the energy efficiency improvement itself changes relative prices (and, thus, real income).” ¹ Adjusting that terminology to the energy storage market is a relatively straightforward exercise in adjusting jargon. Rather than using words like consumer and market that are so common in economic vocabulary, the energy storage community often refers to the same actors as distributed energy resources (DERs) and the grid/ wholesale energy market, wherein “the grid” refers to the host of technologies, platforms and operators that enable the reliable delivery of electricity. As such, applying the indirect rebound effect definition to energy storage begins to clarify the parallels: DERs and the grid respond to the energy demand improvement. Such DER and grid responses are likely to occur because the improvement in consumer energy demand management changes relative reliability scenarios (and, thus, the grid outlook).

Put another way, the theory of indirect rebound effects originally applied to energy efficiency in a fairly straight¬forward manner: Utilizing more energy efficient devices can beget utilization of that device more frequently. By utilizing an efficient device more frequently, the positive effects associated with its more efficient operations are slightly negated. The same can be said of storage: Utilizing energy storage enables more effective utilization of more energy storage devices. But also, by utilizing a single energy storage device across more applications, the benefits associated with its performance become increasingly fuzzy.

Indirect System Effects

At the heart of identifying indirect effects is pinpoint¬ing how the original desired solution affects other components in the system. Identifying the eventual system effects for the deployment of energy storage is still very much an act of gazing upon a crystal ball. However, it is clear that the industry is trending towards increasingly distributed variable generation, and energy storage can help mitigate this variability. Additional variable generation improves the value proposition for energy storage (assuming the cost of the storage is less than traditional infrastructure improvements), making the value proposition for more variable generation (and the additional storage) more attractive (the cycle continues).

The Gridwise Architecture Council is working through what it calls the Transactive Energy Framework. In this scenario, energy resources are constantly negotiating the value of energy locally at the source of the distributed generation as well as an aggregation of distributed resources in the wholesale energy market. The indirect performance effects outlined earlier would be naturally handled by a distribution market that responds in real time to the local factors balancing supply and demand. Rather than each resource serving its own self-interests with little regard to the bigger picture, such a market would necessitate a balancing of both local factors and larger system level considerations.

Summary

Within this piece, multiple effects of disrupting the normal performance of energy storage systems were covered. Brief descriptions of each are below:

• Direct Rebound Effect – The energy storage system returns to higher levels than average immediately following a DR event before returning to roughly average performance.
• Indirect Performance Effect – The energy storage system offsets DR performance by performing differently during other parts of the day.
• Indirect Rebound Effect – Utilization of energy storage across multiple applications reduces the benefits associated with any single application.
• Indirect System Effect – Energy storage transitions the electricity architecture to a new paradigm.

While each of these effects may be weighted differently by different stakeholders, the performance of a distributed storage device does not occur in a bubble. The greater context of the surrounding landscape must be considered. These effects must be built into predictive models as we consider the future of our distribution system. 

¹ Gillingham, K., Rapson, D., Wagner, G., 2015. The rebound effect and energy efficiency policy.