Energy Storage: Designing for Performance


by John Jung and David Miller, Greensmith

Energy storage represents one of the largest opportunities in the U.S. power market, if not the global grid in general. Storage could be a transformative influence on the industry once it becomes integrated into supply chains; think of the profound effects on world trade and regional demographics unleashed by the advent of refrigerated rail cars in the 19th century. Electricity remains the last commodity that is fundamental to our way of life but that has not enjoyed the benefits of economical, distributed and widely available storage.

Energy storage, however, isn’t easy to deliver or deploy. No standard approach addresses all of its complexities, particularly at grid scale. Because grid-scale storage is a complex system of systems technology, we also will not see a rapid convergence on a core technology-like we’ve seen with crystalline silicon in solar-that will enable rapid scaling and a race to commodification. Battery chemistries, such as sodium, lithium and zinc, as well as other storage technologies will continue to evolve for years.

Why? Storage is not simply a battery or a fixed-function point product like photovoltaics or hard drives. Instead, storage systems are multifunction resources tasked with managing different and sometimes conflicting services and applications. A single storage system can be used for energy-centric applications such as peak shifting at one point during a day and then transition to power-centric applications like frequency regulation at another time. Storage systems will become the equivalent of computers for the grid, intelligently coordinating a network of activities on behalf of grid operators, power companies and their customers while juggling revenue streams, regulated markets and forms of grid congestion.

Unlike solar installations where technology design decisions are relatively straightforward and performance degradation is fairly predictable, energy storage system design involves a cascading array of variables that can be complicated further because both battery performance and use-case scenarios likely will change substantially over the lifetime of the system-sometimes up to 20 years. One size will not fit all, and a single battery type never will do the same.

Invariably, many systems will fail to achieve their promised performance specifications and return on investment (ROI) goals because developers or customers ignored the risks and complexities. If mistakes become widespread, the momentum behind energy storage could be stalled for years.

As a result, utilities and others must choose wisely. Let’s examine energy storage design approaches for three distinct use cases: solar photovoltaic (PV) energy time-shifting, frequency regulation and solar PV ramp rate control.

In each case, choices regarding battery selection, the use of intelligent energy management software and the implementation of a smart charging algorithm directly affect the energy storage system’s performance and resulting ROI. Many factors are not obvious at first. Only with careful planning can companies get the most out of storage.

PV Time-Shifting

In this example, a solar developer planned to add energy storage to an existing solar PV system. Located in California, the multimegawatt PV system had a peak energy output that exceeded its off-take capacity and was obligated to curtail output during some hours of the day. By adding energy storage to the PV system, the solar developer sought to time-shift some of the PV output to later hours of the day. By time-shifting some of the energy, the solar developer would be able to reduce PV curtailment and increase revenue from the original off-take agreement.

A technology evaluation helped determine the optimal battery technology to time-shift solar energy output at the lowest possible cost. The four battery options evaluated were:

  1. Li-ion: lithium manganese oxide (LMO) battery;
  2. Li-ion: lithium iron phosphate (LFP) battery;
  3. Lead acid battery; and
  4. Lead acid with Li-ion replacement purchased (in the future at a lower forward cost).

The levelized cost of energy (LCOE) modeling indicated that the lowest LCOE would be achieved using LMO batteries. This result might be surprising because the LMO batteries had the highest upfront capital cost of the battery chemistries examined. The LMO LCOE was lowest, however, because the control software could optimize the battery operation to reduce performance degradation over time compared with other chemistries, allowing the system to be in operation for 15 years. The additional years of energy discharge decrease the project’s LCOE.

On a project basis, LFP batteries and lead acid batteries resulted in higher LCOEs than LMO. A fourth hybrid option in which lead acid batteries were initially installed and then replaced at the end of year four also was evaluated. The benefit of this scenario would be the ability to purchase Li-ion batteries in 2019, when the projected price of Li-ion batteries will have decreased. The LCOE of this hybrid project over a 20-year project life was cheaper than using lead acid alone but was still more expensive than procuring a Li-ion battery initially. Given this modeling, LMO batteries were selected for the project with a plan to use them now until the end of their 15-year useful life as opposed to waiting for prices to drop.

Click to Enlarge

PJM Frequency Regulation

Wholesale frequency regulation markets are one of the most lucrative markets for storage. Federal Energy Regulatory Commission (FERC) Orders 755 and 785 instruct system operators to pay frequency regulation market participants for speed and accuracy-two metrics in which energy storage excels.

Although the opportunity for energy storage is significant, intelligent energy management software can maximize energy storage system returns.

For example, within the PJM regional transmission organization, pay-for-performance frequency regulation has been implemented via two separate regulation signals: RegA and RegD. Resources that respond to signals are paid based on their capacity and performance, as measured by PJM’s performance score.

The RegD signal is designed to be energy-neutral over a 15-minute period, meaning a battery would be called upon to discharge the same amount of energy as it charges over this period. Yet, despite this energy neutrality, all energy storage technologies have round-trip efficiency losses. Because of these losses, the state of charge of a battery will decline over the day, even when responding to an energy-neutral signal.

Some energy storage technology vendors have solved this problem by periodically removing their energy storage system from the frequency regulation market for at least an hour to charge the battery to an appropriate state of charge. This might offer a high performance score during the time of day when the energy storage system is “in market,” but the system cannot participate in the market for two to three hours each day. In some cases, this strategy also might increase degradation of the energy storage system’s batteries, leading to shortened life expectancy.

Intelligent energy management software enables an energy storage system to participate in the frequency regulation market over an entire day without the battery state of charge declining and while maintaining a performance score above 95 percent.

Because the energy storage system can participate in the frequency regulation market over more hours of the day, energy storage system owner returns are increased 8 to 12 percent. Meanwhile, intelligent energy management software controls the battery state of charge to reduce degradation. A reduction of degradation allows an energy storage system owner to initially size a smaller battery system for the same frequency regulation capacity. In one case, proper sizing with intelligent energy management software reduced the quantity of batteries procured 57.5 percent.

Solar Ramp Rate Control

Solar ramp rate control refers to a case in which a solar PV system is limited in the rate of change of energy output it is allowed to send to the grid. PV output fluctuations are common when clouds pass over a solar array. These output fluctuations are a particular problem in remote areas or island grids with high levels of solar PV, where changing solar conditions can cause voltage fluctuations on the grid and damage equipment.

Intuitively, solar ramp rate control is similar to frequency regulation. When a cloud passes over a solar array, solar output drops immediately, and energy storage must discharge to keep the net output stable. When the cloud moves, the solar output resumes, and energy storage must charge to keep net output stable.

Experience shows that when using an energy storage system for solar ramp rate control, energy demands on the battery are heavily biased toward discharge. During the day as clouds move over solar arrays, the energy storage system discharges more than it charges. This bias toward discharge leads to an increase in the energy capacity needed for the energy storage system, thereby increasing cost.

This problem can be addressed if one installs an energy storage system that employs a smart charging algorithm. The solar ramp rate control smart charging algorithm will ensure that during stable solar output the energy storage system’s battery is kept at a sufficient state of charge such that it is not depleted during the day. By dynamically charging the battery based on weather conditions, the system can reduce the total size of the battery required for the system. For example, the use of smart charging algorithms on one ramp-rate control system allowed the required capacity of the energy storage system and resulting capital expenditure on batteries to be reduced 50 percent.


Battery choice, the use of intelligent energy management software and the implementation of smart charging algorithms all can affect energy storage system performance and cost significantly. Failure to understand the nuances of battery behavior can result in a choice of battery with a higher LCOE than potential alternatives. Lack of intelligent energy storage software can lead to suboptimal energy storage system performance in frequency regulation markets, limited market participation and increasing battery degradation. Smart charging algorithms can reduce energy storage system capacity requirements for ramp rate control applications, lowering overall battery capital expenditures.

Variety will continue to be the norm for grid-scale storage systems because of varying use cases. Developers will converge on best practices for distributed and grid-scale storage but invariably recognize the need to customize their deployments. An analogy can be made to data center architecture: All data centers, generally, perform the same task, i.e., store or transmit data for many users, but no two are identical and no two are used in exactly the same way. In the same way, energy storage systems might be built out of commodity components, but they will be distinct. In both cases-data centers and energy storage systems-careful evaluation and planning are essential to get the most out of the investment.

John Jung is the CEO of Greensmith, a provider of energy storage management software and associated services. He has spent the past six years focusing on optimizing storage for utility customers. Reach him at

David Miller is director of business development at Greensmith. He has written extensively on the energy industry. Reach him at

More PowerGrid International Issue Articles
PowerGrid International Articles Archives
View Power Generation Articles on
Previous articleDistribuTECH 2015 Blows Past Records
Next articleStaying Ahead of Renewables

No posts to display