2022 BATTERY

SCORECARD
2022 Battery Scorecard

CONTENTS

1 Introduction ____________________________________________
03

2 What is the Battery Scorecard? ________________________ 05

3 How to interpret the Battery Scorecard ___________________

07

4 Trends in energy storage _____________________________ 12

5 Scorecard results ________________________________________

16

5.1 Battery cells ___________________________________________

17
5.2 Battery Management System (BMS) ______________________
21
5.3 Safety ________________________________________________
22

6 2023 and beyond ____________________________________________

23

2
2022 Battery Scorecard

1 INTRODUCTION

3
2022 Battery Scorecard

1 INTRODUCTION
Energy storage is a key technology enabling the energy transition. Over the past ten years
we’ve seen significant growth of utility-scale and behind-the-meter stationary storage along
with electric vehicle adoption. These markets are built primarily upon a single technology:
the lithium-ion battery.

While increases in battery costs have slowed growth this year, For over a decade, DNV has been helping energy storage
we believe the market will quickly recover and eventually partners identify these types of uncertainty to help manage
achieve over 4,500 cumulative GWh of stationary energy and reduce risk. Every release of our Battery Scorecard shares
deployments by 2050 (DNV's Energy Transition Outlook). new insights from the data gathered during testing, analysis,
Energy storage—due to its versatility, novelty, and and forecasting to help characterize asset performance.
complexity—make it a very interesting and exciting market, This 4th edition of DNV’s Battery Scorecard incorporates an
which will continue to grow in the coming years as it enables interactive component through an online dashboard that
us to cost effectively decarbonize the grid and automotive provides deep insights into battery testing conducted at the
sectors. Battery and Energy Storage Technology (BEST) Test &
Commercialization Center (BEST Test Center) in Rochester,
Battery storage has proven valuable across multiple global New York, and free limited access to our Battery AI forecasting
markets, taking a similar trajectory in each. It typically is first tool.
introduced to reduce peak demand for commercial and
industrial applications and to provide ancillary services to Battery storage—in both stationary energy storage systems
balance the grid due to its fast-responding and flexible and electric vehicles—has an important role in accelerating the
nature. As markets evolve, we see energy storage systems uptake of renewable generation and decreasing greenhouse
providing clean sources of capacity to the electric grid and gases. We at DNV look forward to helping lead the energy
shifting renewable energy from periods of high supply to transition and would love to hear from you on how we can
times of high demand. help achieve this together.

Battery energy storage systems are adopted when their value

exceeds costs, which is becoming common in many areas.
Energy storage system costs are typically dominated by the
battery, which has seen a recent uptick in price due to lithium
and other raw material shortages. This price volatility has
created uncertainty in the energy storage market.

4
2022 Battery Scorecard

2 WHAT IS THE BATTERY

SCORECARD?

5
2022 Battery Scorecard

2 WHAT IS THE BATTERY SCORECARD?

DNV’s Battery Scorecard is a free, publicly available report and online dashboard that illuminates
on some of the most pressing questions around batteries:
• Who are the major battery suppliers, and have they been vetted?
• How do batteries degrade?
• What is a battery’s useful life?
• Are some batteries safer than others?

These are complex questions that require a comprehensive

understanding of battery technology, system integration and DNV'S NEW BATTERY SCORECARD ONLINE
control, testing strategies, manufacturing, and the energy DASHBOARD
storage market. The Battery Scorecard and online dashboard Are you wondering why the Battery Scorecard now
address these questions head-on and are discussed in detail incorporates an online dashboard? Our insights are
in the following sections. based on considerable amounts of data and DNV has
incorporated advanced computational methods to
There are many current events that raise questions about the process and convert this data into meaningful insights.
energy storage market, including COVID-19, the Russian We couldn’t share everything in a static report,
invasion of Ukraine, and mineral supply. We address these so the online dashboard now gives you the ability to
uncertainties and how they affect the market in Section 4. dynamically filter and focus on what you care about
most. The dashboard also links to some of the most
advanced online modeling tools on the market,
including DNV’s predictive degradation tool Battery AI.
DNV plans to share updates more often and will be
updating the online dashboard with key insights
throughout the year, so please check back regularly.

6
2022 Battery Scorecard

3 HOW TO INTERPRET THE

BATTERY SCORECARD

7
2022 Battery Scorecard

3 HOW TO INTERPET THE BATTERY

SCORECARD
The Battery Scorecard provides insights into technology readiness, degradation, useful life, and
safety. DNV’s approach to assessing these topics is detailed below.

Technology readiness/bankability Many battery suppliers have limited global exposure or are
Selecting suitable technology suppliers can be one of the new to stationary energy storage, so it can be challenging to
most important and challenging endeavors for product make informed purchasing decisions.
designers, system integrators, and project developers.
While batteries have been commercially available for Here is our recommended approach when evaluating
decades, most stationary energy storage products are less vendors within the energy storage market:
than five years old.

1 START WITH THE CELL

Battery energy storage systems all have battery cells as their core component, which dictate performance, safety, and cost.
Find out who manufactures the cell (if different from the system integrator), and how long that cell has been in production.
DNV shares information about cell manufacturers of various battery energy storage systems in the online dashboard to help
you better understand these products. This information can be found in the Bankability section of the dashboard.

2 EVALUATE TOTAL DEPLOYMENTS OF CELLS AND SYSTEMS

Total deployments of the underlying battery cell and the integrated system are both good indicators of how reliable the
product will be. There are often many bugs to work through after product launch. More experience in the field often leads to
improved products. DNV shares the top battery cell providers globally below in Section 5.1.1 and in the online dashboard.
This information can be found in the Battery Cell Market Overview section of the dashboard.

3 REQUEST THE 'BANKABILITY REPORT'

Independent vetting of the integrated system is critical to make an informed decision. Battery manufacturers should have
Bankability Reports on their products and should be excited to share these with potential customers to prove they have been
independently vetted. DNV has reviewed many products on the market spanning battery cells, residential and utility scale
battery energy storage systems, controls, integrated dc-coupled solar + storage systems, and non-lithium storage systems,
with the summary list shared in the online dashboard. This information can be found in the Bankability section of the
dashboard.

8
2022 Battery Scorecard

4 REQUEST INDEPENDENT TEST DATA TO VALIDATE PERFORMANCE, WARRANTY CLAIMS, AND SAFETY

Testing provides validation of each product’s design, operation, and safety. Independent testing verifies that the product will
operate as specified without manufacturer intervention. DNV recommends requesting independent test data from all battery
suppliers before making a purchasing decision to better understand the product and its impact on your project.

DNV provides a list of battery cells that have been independently performance tested and details use-case specific results
in Sections 5.1.2 - 5.1.5 and in the online dashboard. This information can be found on the Battery Performance section of the
dashboard. The results suggest early degradation trends and preferred operating parameters for each product. Third party
safety testing is also required and should be provided by manufacturers for battery cells, integrated modules, and full systems.
Burn testing is also required to understand the fire and explosion risk associated with battery products. These safety risks are
discussed below and further in the online dashboard. This information can be found on the Battery Safety section of the
dashboard.

After completing initial vetting of battery suppliers discussed operation. While most testing is conducted for 6 to 12
in steps 1-4, you can now progress to vendor selection and months, early failure modes and degradation trends can be
more accurately predict degradation, augmentation needs, identified in the first month or two of testing, especially in
useful life, and battery safety. stress cases. Many are also pushing for longer testing to
better understand when a battery cell fails, entering a period
Battery degradation of accelerated, non-linear degradation and ultimately losing
Battery degradation can only be determined through cell all capacity to hold a charge. DNV shares Battery Scorecard
testing, which DNV conducts at the BEST Test Center in Test results from the BEST Test Center and from Witness
Rochester, New York. DNV also collects data through onsite testing in the section below and in the online dashboard
witness testing conducted in the labs of cell manufacturers These results highlight initial trends in performance across
around the globe. While battery manufacturers also conduct key categories.
their own testing, these test results are not currently
incorporated into the Battery Scorecard because the test Additionally, there is a separate form of degradation that
methods and results have not been independently verified. occurs independent of use, often called “calendar fade”.
DNV categorizes test data as follows: This phenomenon may be accelerated at elevated
temperatures and SOC, depending on the cell, and should
• DNV testing: testing conducted at the BEST Test Center be a prime focus of project developers when calculating
by DNV employees using the Battery Scorecard Test Plan. long-term degradation, planning augmentation, and
• Witness testing: testing overseen by DNV at a battery scheduling battery delivery and commissioning. Multi-month
manufacturer’s site and using Battery Scorecard Test Plan. delays could have serious impacts on initial energy capacity
• Manufacturer testing: not included in the Battery of your storage system and may or may not be covered in the
Scorecard since test methods and results have not been product warranty. DNV shares Battery Scorecard Test results
independently verified. from the BEST Test Center and from witness testing below
and in the online dashboard in the Calendar Fade section.
All testing follows DNV’s standard testing protocol known as
the Battery Scorecard Test Plan, which evaluates battery cells Finally, to encourage even more independent testing, DNV
across four primary stress factors: charge and discharge rate will publish a Recommended Practice that standardizes
(C-rate/P-rate); state of charge (SOC) swing (0 >100 > 0 =100, battery cell performance testing. The goal is to help guide
25 > 75 > 25 = 50, etc.); average SOC; and ambient the industry towards a standard best practice for battery cell
operating temperature. The Battery Scorecard Test Plan testing so that data across chemistries, form factors,
includes nearly 40 separate tests that cover typical operating manufacturers, and from different labs can better be used
windows and stress cases for stationary and mobile battery interchangeably.

9
2022 Battery Scorecard

Useful life These sample use cases help demonstrate the expected
Useful life is the period of operation when a battery energy degradation of each BESS and can be used by project
storage system (BESS) can predictably charge, store, and developers and lenders to compare against warranties being
return energy for a useful application. When a battery provided by battery manufacturers and integrators.
operates beyond its useful life, it degrades unpredictably and
loses its capacity to store energy at an accelerated rate until A battery’s useful life is highly dependent on use case, as
it is unable to hold a charge. The useful life of a BESS is discussed above, so manufacturers and integrators
dependent upon the underlying useful life of each battery typically limit daily and/or annual usage (known as
cell, with the combined performance being dependent on “throughput”) so that the useful life can be converted into
how the cells are integrated, operated, maintained, and years, which is needed when developing a maintenance and
balanced within the BESS. If individual cells degrade replacement schedule for a project. While stationary BESS
differently, it can severely limit the system’s energy capacity, typically have warranties ranging from 10 to 25 years, these
or worse, it could cause accelerated heating and eventual warranty periods are typically achieved by adding batteries
failure of the cell. to, or “augmenting”, the system during operations. There are
very few (if any) stationary batteries on the market today that
The transition from predictable (generally linear) to have been operating for 15+ years, even though many new
accelerated degradation is often referred to as the “knee” storage projects have warranty periods that exceed this.
or “shoulder” in the degradation curve and should be The useful life of an individual battery may be much less
avoided. Predicting the knee in a BESS degradation curve is than the warrantied period, depending on application.
very difficult, given the multiple factors involved in DNV considers the useful life of individual batteries to be
integrating and controlling the batteries to achieve 10 to 20 years or when the batteries have degraded to
project-specific requirements. DNV has developed an 60-65% of initial energy capacity, whichever comes first.
advanced software modeling tool called Battery AI to predict
project-specific degradation and prevent BESS operators Vehicle battery warranties are typically based on distance
from hitting the knee. Battery AI imports data from Battery traveled (i.e., 200,000 km) and duration (i.e., 10 years), which
Scorecard testing and implements battery-specific is one step removed from the factors that cause degradation
degradation algorithms to create a “digital twin” of a battery (such as throughput and calendar fade). Typical automotive
cell that can predict cell and system performance over a use cases differ widely, with passenger vehicles having
range of user-defined cases. much more variability than an electric bus fleet with more
predictable routes and charging profiles. DNV considers
To demonstrate how useful life modeling in Battery AI is 6- to 12-year warranties for vehicle batteries to be standard,
conducted, we selected four representative use cases to with energy capacity limits set at 80% of initial capacity (rather
analyze BESS performance, including: than 60-65% in stationary applications).

Marine electrification includes additional risk factors that can

• Firm frequency response in the United Kingdom: accelerate battery degradation, such as water and salt
A one-hour battery participating in the Firm infiltration, and has rigorous safety standards given the risk
Frequency Response market in the United Kingdom to life and property if failures occur at sea. Emerging aviation
in 2018; applications demand even more reliability and safety.
• Merchant storage in Texas, U.S. (ERCOT):
A two-hour commercial battery providing energy and The Battery Scorecard’s online dashboard links to the Battery
ancillary services products in the ERCOT market; AI service and provides a demonstration of the tool’s useful
• Solar firming in a moderate climate (25°C): life prediction capabilities. All registered users can sample
A four-hour battery co-located with a solar farm to Battery AI’s functionality. Customers can also purchase
shift production to peak price hours under a detailed cell modeling in Battery AI to pre-screen different
moderate temperature; cells and see which are best for specific applications. DNV
• Solar firming in a hot climate (40°C): expects this tool to get increased focus from developers and
A four-hour battery co-located with a solar farm to lenders as useful life, and in turn project cashflow, receive
shift production to peak price hours operating in an increased scrutiny as the market expands.
elevated temperature environment.

10
2022 Battery Scorecard

Battery systems aren't worth installing

if they can't be run safely

Battery safety
Arguably the most important aspects of the design are
safety features, and, like performance, safety features need
to be vetted thoroughly. There are many required safety
certifications for battery cells, modules, energy storage
systems, and electric vehicles across regions, including IEC,
UL, and SAE Standards, to name a few. Safety of an energy
storage system builds up from the battery cell and is relevant
to every stage of a product’s lifecycle, based upon both code
requirements and best practices. Battery safety has quickly
become one of the most central focuses in evaluating battery
products and projects due to some high-profile battery fires
that have made news globally.

DNV includes anonymized, aggregated UL 9540A burn

testing results in this year's Battery Scorecard to draw focus
to some of the most important questions being asked related
to battery storage safety. In the section below and within the
online dashboard, we address key themes to consider when
evaluating battery cells and BESS for safety considerations.

11
2022 Battery Scorecard

4 TRENDS IN ENERGY STORAGE

12
2022 Battery Scorecard

4 ENERGY STORAGE MARKET

OVERVIEW
Many factors affect the energy storage market, with some underlying drivers providing
predictable growth trends. The ability of an energy storage system to stabilize the electricity
grid and shift daytime solar energy into morning and evening hours has been a consistent
driver of growth.

In general, early markets form in new global regions around Li-ion chemistries dominate
ancillary services, which balance the electricity grid when After over a decade of investments into electric vehicles,
consumption (load) is not perfectly aligned with consumer electronics, and more recently stationary energy
generation. Ancillary services require fast-responding storage, Li-ion batteries have become the dominant battery
resources to perform “grid support” such as frequency and on the market. Within the broad category of Li-ion batteries,
voltage control and add power capacity, for which BESS are nickel manganese cobalt oxide (NMC) chemistry has roughly
particularly well suited. Other early battery energy storage 50% of the market, with lithium iron phosphate (LFP) and
success stories are found in commercial-and-industrial (C&I) nickel cobalt aluminum oxide (NCA) chemistries gaining
behind-the-meter applications, such as demand charge quickly. There has been regional focus across battery
reduction, where batteries provide energy during periods of chemistries, with many leading NMC manufacturers based in
high pricing. This behind-the-meter application reduces peak Korea and LFP manufacturers based in China, though some
power costs, thus saving money for the C&I facility owners. notable manufacturing development in Europe and the U.S.
These types of systems can also provide resilience, such as is expected to compete in the coming years. The big story is
temporary backup power during a grid outage. that LFP battery manufacturer CATL has become the global
leader in Li-ion battery supply, overtaking the long-standing
Other factors affecting the energy storage market are less NMC manufacturer LG ES, which now holds the second spot
predictable, where reliable information is hard to come by (see Scorecard Results below for a list of top 10
and reporting seems to change daily. Government mandates manufacturers). Stationary storage is also trending toward LFP,
and incentives, commodity mineral pricing, battery supply while electric vehicles still primarily employ NMC and NCA
shortages, new disruptive technologies, and both regional due to their higher energy densities. For the near term, Li-ion
and global politics create uncertainty in the market. Trends in both transportation and stationary energy storage should
across these less-predictable categories are discussed next. remain dominant for at least the next three to seven years.

Next-generation technologies, such as silicon, sodium, and

lithium metal anodes, solid-state electrolytes, new cathode
material and cell manufacturing processes, flow batteries, and
other non-lithium technologies, could play an important role
in enabling these price reductions, which is discussed further
in Section 6.

13
2022 Battery Scorecard

Current price volatility COVID-19 challenged storage sector growth in some

In 2010, average Li-ion battery pack prices, across different countries, but deployment in the U.S. has been strong, with
battery end uses, were above $1,200 per kilowatt-hour. new hybrid solar + storage driving growth in sunnier regions.
These prices have fallen 89% to $132/kWh in 2021, according China will compete with the U.S. for greatest volume of
to Bloomberg New Energy Finance (BloombergNEF Annual deployment of lithium batteries in the coming years—and
Battery Price Survey 2021), and this is a 6% drop from Chinese suppliers of lithium batteries are poised to provide
$140/kWh in 2020. the lowest-cost products. China is set to be the world leader
in storage capacity by 2024.
But the end of 2021 and first half of 2022 has seen a sharp
uptick in costs, driven primarily by the rising cost of lithium Battery safety testing: critical yet complicated
and other raw materials, global supply constraints, broader With European, North American, and Asian energy storage
inflation, and production curbs in China. While the coveted markets quickly growing and tenders being a race to bottom
$100/kWh battery is still attainable in the longer-term, battery on price, it is more important than ever that owners and
costs are significantly higher than this in many regions, with operators consider battery safety. Relatively rare but high-
project developers in North America getting quotes from profile battery fires such as the Victoria (Australia) Big
battery manufacturers up to nearly $500/kWh in mid-2022. Battery fire in July 2021 and London Buses in May 2022 bring
Average battery costs in 2021 were lower, though they saw increased scrutiny to the deployment of future EVs and
an imbalance across regions, with the lowest costs in China, stationary systems. The industry needs to learn—and is learning—
at $111/kWh. Packs in the U.S. and Europe cost 40% and 60% from these events, with public release of root cause analyses
higher, respectively, according to BNEF. Prices have fallen as becoming more common. But with wider deployment the
the adoption of LFP has increased and as the use of industry must continue to reduce failures to avoid widespread
expensive cobalt in nickel-base cathodes has lessened. loss of equipment, environmental impact, or worse, and in turn
On average, LFP cells were nearly 30% cheaper than NMC lose confidence from consumers and regulators.
cells.
For stationary storage applications, results from industry
This volatility has impacted project financials, and the market standard test methods such as UL 9540A (Test Method for
is already responding. In the U.S., which is a leading market Evaluating Thermal Runaway Propagation in Battery Energy
for grid energy storage, the market is being driven by new Storage Systems) are critical to obtaining approval from
state-level storage mandates and supportive federal policy permitting authorities. UL 9540A is a test method, not a
such as Federal Energy Regulatory Commission Order 2222. standard by which a battery passes or fails the test. Instead,
The U.S. energy storage market is expected to expand from test results provide key information to inform system design,
an annual deployment of 5 GW/14 GWh in 2022 to spacing, siting decisions, and emergency response plans.
14 GW/50 GWh in 2026, according to Wood Mackenzie The method includes a progression of testing from individual
projections. Off-takers are renegotiating (and paying more cells to modules, units, and installations, shown in Figure 4-1.
for) purchase agreements to better reflect the cost of
procuring energy when they need it.

21 mm
18 mm

65 mm 70 mm

18650 cell 2170 cell

Cathode Separator
Anode

24 Tray rack
Header Can

Figure 4-1 UL9540A testing is conducted at the cell, module, unit (racks), and installation (system) level

14
2022 Battery Scorecard

Figure 4-2 Cell-level testing helps understand failure modes and thresholds of each battery cell

Cell-level testing evaluates fault conditions and failure modes Note the variation in temperatures at different measurement
at the smallest divisible level of a battery system—the cell. The points on the cell—up to roughly 90°C difference at venting,
photos in Figure 4-2 show an example UL 9540A prismatic and 110°C difference at thermal runaway. The test method
cell test setup, with the cell sandwiched within pressure plates is not always clear which temperature measurement points
and thermocouples applied to measure temperature (at left), should be used in reporting the results, causing confusion
and an example of a cell's response to nail penetration (at when being used to inform downstream design and safety
right). Note that actual 9540A cell-level tests occur within an decisions. DNV recommends conferring with a battery testing
enclosed environment. expert when interpreting UL 9540A test results.

The plot in Figure 4-3 is from an example cell-level 9540A

500
overheating test, showing temperatures from the start of the
test through about 60 minutes, when thermal runaway occurs, Thermal
runaway
and the onset of subsequent cooling. The thin grey line 400
Gas
represents the atmosphere within the test chamber. The three venting
colored lines represent different measurement points on the 300
Temperature (°C)

battery surface, including front center (red), side center (blue),

and top (green), with only the front and side being in direct 200
contact with the heater. During the test, heat is applied by an
external film heater wrapped around the cell to raise the cell’s
100
temperature by 5°C per minute. At about 40 minutes into the
test, liquid within the cell’s electrolyte boils and the gases
vent, leading to short-term cooling of the battery (dips in blue 0
0 10 20 30 40 50 60 70 80
and green lines) and an increase in the chamber’s ambient Time (minutes)

temperature (grey line blip). Continued heating leads to

Chamber ambient Cell front center Cell side center Cell top
thermal runaway of the cell at about 60 minutes—when the
temperature increase from reactions within the cell exceed
Figure 4-3 Cell-level UL 9540A test results show temperatures
the temperature increase from the external heater. During
rising steadily, dipping, rising again, and then spiking during thermal
thermal runaway, all temperatures increase rapidly. runaway

15
2022 Battery Scorecard

5 SCORECARD RESULTS

16
2022 Battery Scorecard

5 SCORECARD RESULTS
DNV recently tested 19 battery cells through the Battery Scorecard Testing program and
includes findings from these tests below. In some cases, the cell manufacturer has agreed to
share its name, though many of the manufacturers chose to remain anonymous.

In the categories where the manufacturers chose to remain

5.1 Battery cells
anonymous, DNV provides general insights to guide
downstream partners in evaluating and selecting batteries
5.1.1 CELL MANUFACTURING VOLUME
for their application. DNV plans to release additional results
The top 10 battery cell manufacturers by volume for 2022 are
as they become available. We also hope to share additional
projected in Table 5-1. These results include all battery cells
names of currently anonymized manufacturers as these
produced across EV and stationary energy storage systems
manufacturers see the value in transparency.
(ESS). With only a few notable exceptions, most battery cell
manufacturers have >90% of their cells going to EVs.
Beyond performance testing results, DNV is also sharing
Other battery cell manufacturers not listed in the top 10 are
publicly available information on battery cell manufacturing
grouped together in the “Other (cumulative)” category,
and system deployments, where the cell manufacturers and
totaling 235.1 GWh of projected cells produced in 2022.
system integrators have released this information publicly
online.
TOTAL 2022 CELL
MANUFACTURER
PRODUCTION (GWh)
For battery safety results, all manufacturers have remained
anonymous; therefore, we have focused on presenting 1. Contemporary Amperex
general trends and comparisons between chemistries rather Technology Co Ltd (CATL) 132.0
than highlighting specific manufacturers or products.
2. LG Energy Solution 93.9
DNV evaluates cells and systems across the following
categories: 3. Panasonic Corp 60.1

4. BYD Co Ltd 58.6

• Cell manufacturing volume
• Cell performance: <2-hour grid support services 5. Samsung SDI 47.1
- LFP category
- NMC category 6. SK Innovation 32.0
• Cell performance: 4-hour solar shifting
7. TianJin Lishen Battery
- LFP category 21.9
Joint-Stock Co Ltd
- NMC category
• Cell performance: high-power vehicle application 21.5
8. Gotion High Tech Co Ltd
- NMC category
• Calendar fade 9. EVE Energy 18.5
• Battery management system optimization
• Safety: offgas and thermal runaway temperature 10. Amperex Technology Ltd (ATL) 17.5
thresholds
• Safety: offgas composition Other (Cumulative) 235.1

Table 5-1 Leading cell manufacturers by 2022 projected volume

The results for each category are summarized next. Data from Benchmark Mineral Intelligence

17
2022 Battery Scorecard

5.1.2 CELL PERFORMANCE: <2-HOUR GRID SUPPORT The graphs in Figure 5-1 show cell testing results across a
Out of 19 cells included in this year’s Battery Scorecard, range of test parameters for each chemistry, including low,
the top 3 performing LFP and NMC battery cells within the high, and room temperature conditions, a range of SOCs, and
<2-hour grid support category are presented here. Battery C-rates of 0.5 and 1. Generally, capacity degradation happens
Scorecard Testing that evaluated 0.5C-1C performance across more quickly in the first year (assuming 365 equivalent full
various temperatures were included in these results. cycles per year), falling 3% to 5% in the first year before
leveling out to an annual degradation rate between 1% and
3% per year, depending on use case, cell type, SOC, and
TOP TOP
temperature. Cell operation at the upper and/or lower
PERFORMERS PERFORMERS
regions of each category, even within specified windows, such
(LFP) (NMC)
as hot or cold temperatures, high C-rates, or high SOC
thresholds, can result in capacity degradation at 8-10% per
year.
CATL Narada OEM-6 OEM-2 OEM-3

LFP Performance NMC Performance

Product model (generic) CATL-1 CATL-2 LFP-2 Narada-1
Product model (generic) NMC-5 NMC-6 NMC-7

100 100
Normalized capacity (%)
Normalized capacity (%)

90 90

Manufacturer Manufacturer
80 80
CATL OEM-2
Narada OEM-3
OEM-6

70 70
0 500 1000 1500 2000 0 500 1000 1500 2000

Turnover (equivalent full cycles) Turnover (equivalent full cycles)

Figure 5-1 Top 3 performing cells in the <2 hour grid support category

18
2022 Battery Scorecard

LFP Performance NMC Performance

Product model (generic) CATL-1 CATL-2 LFP-2 Narada-1 Product model (generic) NMC-5 NMC-6 NMC-7
100 100
Normalized capacity (%)

Normalized capacity (%)

90 90

Manufacturer Manufacturer
80 CATL 80 OEM-2
Narada OEM-3
OEM-6

70 70
0 500 1000 1500 2000 0 500 1000 1500

Turnover (equivalent full cycles) Turnover (equivalent full cycles)

Figure 5-2 Top 3 performing cells in the 4-hour solar shifting category

As discussed above, each cell eventually falls over a knee and

begins to degrade rapidly, after which the cell has exceeded TOP TOP
its useful life. This is demonstrated in Figure 5-1 by the lower PERFORMERS PERFORMERS
light blue curve for LFP cells. These two cells performed (LFP) (NMC)
admirably compared to their peers for use cases that aligned
with their design specifications but demonstrated early and
rapid degradation when stressed under a higher than CATL Narada OEM-6 OEM-2 OEM-3
designed C-rate and exposure to temperature extremes.
Ideally, all cells would be tested to the end of their useful
life, or when they reach the knee; however, gentler cycling
conditions and improved cell performance has extended the In contrast to the grid support use case above, none of the
testing timeline to reach this point. cells in the 4-hour solar shifting use case shown in Figure
5-2 reached the knee of the degradation curve in which the
Shorter duration applications that require <2 hours of energy capacity degradation accelerates asymptotically towards an
storage capacity are typical for grid support, with example end-of-life. There appears to be only minor differences
markets being Firm Frequency Response in the United between LFP and NMC performance, with a more
Kingdom and ancillary services in Texas, U.S. These pronounced differentiation at higher cycles. While only one
applications often require the system to charge and LFP cell and two NMC cells made it to four years of simulated
discharge, which means that the system is held near the daily cycling (~1,500 equivalent full cycles), all three main-
middle of the SOC range so that it can perform both tained nearly 90% of their initial capacity. For the cells that
functions. have not been tested through as many cycles at this condi-
tion, some were trending higher (with lower degradation) and
5.1.3 CELL PERFORMANCE: 4-HOUR SOLAR SHIFTING others trending lower (with higher degradation). Generally,
Out of 19 cells included in this year’s Battery Scorecard, the the 4-hour use case is less impactful on cell degradation.
top 3 performing LFP and NMC battery cells within the 4-hour Lower C-rates tend to produce less heat at the cell level,
solar shifting category are presented here. Battery Scorecard requiring less heat rejection in the thermal management
Testing that evaluated 0.25C performance across various system and allowing for more optimal temperatures.
temperatures were included in these results.
The 4-hour use case is common in U.S. markets, with capacity
products such as resource adequacy in California requiring
four-hour durations. Four-hour applications also favor
coupling with solar since peak generation (midday) can be
shifted to peak demand times later in the day.

19
2022 Battery Scorecard

5.1.4 CELL PERFORMANCE: VEHICLE APPLICATION 5.1.5 CALENDAR FADE

Out of 19 cells included in this year’s Battery Scorecard, the Of the 19 cells included in this year’s Battery Scorecard,
top performing NMC battery cell in the EV category is nearly all cells showed degradation due to calendar fade.
presented here. Battery Scorecard Testing that evaluated 1C Battery Scorecard Testing evaluated these cells by charging
to 5C performance across various temperatures were them to specified SOCs, and then held them at target
included in these results. temperatures without cycling. These tests are intended to
evaluate the impact of SOC and temperature on calendar
fade degradation. The cells were periodically recharged and
then fully discharged to determine their remaining capacity,
TOP PERFORMER OEM-1 with room temperature results presented in Figure 5-4.
(NMC)

Calendar fade

Higher C-rate testing allows for more equivalent full cycles to Product model (generic) LFP-1 LFP-2 LFP-3 NMC-1 NMC-14 NMC-2 NMC-3
100

Normalized capacity (%)

be examined in a shorter calendar period because each cycle
takes less time to simulate (0.25C takes at least eight hours
for a full equivalent cycle while 2C takes only one hour). 95
As shown in Figure 5-3, NMC-1 demonstrates ~90% capacity
retention over 4,000 equivalent full cycles of testing,
representing about 11 years of daily use. It should be noted 90

that other than commercial vehicles (e.g., ride sharing or 0 50 100 150 200 250 300 350

Days
buses), most passenger vehicles do not experience use cases
represented by a daily full equivalent cycle as the majority of Figure 5-4 Calendar fade degradation at room temperature across
vehicle usage in the U.S. is below 40 miles/day. Under more various cells and at 30% and 100% SOC
aggressive test conditions, cell capacity dropped to 65% after
~3,200 equivalent full cycles, which shows the sensitivity of Calendar fade ranged from 1%-4% per year across the cells
these cells to key operational parameters. tested at room temperature and various SOCs.

Temperature had a clear impact on calendar fade, with

NMC Performance
nearly all cells having increased degradation at elevated
Manufacturer
temperatures and decreased degradation at lower
Product model (generic) NMC-1
100
OEM-1 temperatures. Temperature impact on calendar fade
degradation is shown in Figure 5-5.
Normalized capacity (%)

90
Temperature dependence - calendar fade

Product model (generic) LFP-1 LFP-2 LFP-3 NMC-1 NMC-14 NMC-2 NMC-3
80 100
Normalized capacity (%)

70
95

0K 1K 2K 3K 4K 5K

Turnover (equivalent full cycles) 90

0 10 20 30 40 50 60 70

Temp (°C)
Figure 5-3 Top performing cells in the EV/high C-rate category

Figure 5-5 Temperature dependence of calendar fade

degradation across various cells

20
2022 Battery Scorecard

5.2 Battery Management System (BMS) Capacity retention with reduced upper voltage limit
Capacity retention at 10 years/3650 cycles (%)

100%
A well-tuned BMS can save the day (or year)
Battery cells are typically assembled into modules or racks 80%
with an integrated battery management system (BMS).
The BMS controls upper and lower voltage limits of the cells 60%

as they charge and discharge, among other things.

Manufacturers can tune their BMS to allow for wider voltage 40%

limits (more aggressive) to capture more energy per charge/

20%
discharge cycle. They can also set narrower voltage limits
(more conservative) to avoid upper and/or lower charge 0%
states with the goal of prolonging battery life. All cells 0 -0.1 -0.2
Upper voltage limit compared to rated voltage (delta V)
degrade rapidly if operated outside of their preferred voltage
range, so BMS tuning is critical to optimize the tradeoff of
maximizing short-term capacity while not unduly sacrificing Normalized capacity vs turnovers
long-term performance. 100

In Figure 5-6, multiple cells were tested at 100% rated

Normalized capacity (%)

90
capacity, getting charged (up) to the rated upper voltage limit
and discharged (down) to the rated lower voltage limit.
80
Simultaneous tests used the same cell types with reduced
upper voltage limits, by charging to (a) rated voltage,
(b) 0.1 volts below rated, and (c) 0.2 volts below rated. 70

DNV plans to incorporate BMS rankings in future iterations of 60

0K 1K 2K 3K 4K 5K 6K
the scorecard.
Turnover (equivalent full cycles)

Figure 5-6 Degradation rates can be significantly affected by

adjusting the upper voltage limit in the BMS

21
2022 Battery Scorecard

5.3 Safety Gas venting and thermal runaway temperatures

Batteries subjected to overheating, short circuits, or internal Gas venting temp (°C) Runaway temp (°C) NOTES:
• Large differences between venting & runaway

faults can vent gases and, in some cases, reach thermal temperatures are desirable
• Higher temperatures are gemerally desirable

runaway. Understanding the temperatures at which venting 300 • LFP7 shows high temperatures and ~150°C difference
• NMC1 and LFP1 show lower temperatures and almost

and thermal runaway occur, and the composition of the vent no difference between venting and runaway

gases, are important to safe design and operation. 200

5.3.1 THERMAL RUNAWAY TEMPERATURE THRESHOLD

AND DIFFERENTIAL BETWEEN VENTING AND THERMAL 100

RUNAWAY TEMPERATURES
Two temperatures are most important when considering
thermal runaway. The first is the cell venting temperature, 0
LFP7 LFP4 NMC1 NMC2 LFP2 LFP1 LFP3 LFP5 LFP6
or the temperature at which the buildup of gases within the Cell sample
cell are released through the cell’s pressure release vent to
Figure 5-7 Venting and thermal runaway onset temperatures for
avoid rupturing the cell’s casing. The second temperature is
various cells across chemistries
the onset temperature at which thermal runaway occurs.
This is the “point of no return” for the cell, where
uncontrollable self-propagating reactions involved in thermal Typical off-gas compositions include: hydrogen (H2), carbon
runaway begin. While cells may ultimately reach temperatures dioxide (CO2), carbon monoxide (CO), and a variety of
as high as 1000 °C, the temperature at which these reactions hydrocarbons (HxCy). Hydrogen, carbon monoxide, and
begin helps determine monitoring and control mechanisms hydrocarbons are flammable gases that can contribute as a
needed to avoid thermal runaway altogether. While venting fuel source for a fire. If not burned as it is emitted, hydrogen
and thermal runaway of an isolated cell is concerning by itself, contributes to the explosivity of the off-gas; the more
a key design and control mechanism within battery systems hydrogen present, the more energetic the explosion can be.
lies in preventing cascading effects of one cell’s thermal Other hazard considerations for the gas composition include
runaway causing other cells to also reach thermal runaway. carbon monoxide as a toxic gas and carbon dioxide as an
asphyxiant.
As is evident in Figure 5-7, cells from different manufacturers
and chemistries have vastly different venting and thermal The volume and ratio of gases is largely dependent on the
runaway onset temperatures. This data from UL 9540A materials used inside the cell such as the cathode, electrolyte,
cell-level tests had thermal runaway initiated using a standard and anode. A cell of the same chemistry and size might have
repeatable methodology. It is generally considered favorable a different composition ratio due to the other components
to have a higher degree of separation between the gas within the cell. The average off-gas composition from UL
venting and thermal runaway onset temperatures, especially if 9540A data collected for a number of different cells across
gas detection is available within the energy storage system. chemistries and manufacturers is shown in Figure 5-8.
If gas is detected early, there is more time to catch an
overheating cell and prevent thermal runaway from Average off-gas composition
occurring altogether. If the venting and runaway tempera- Lower flammability limit (avg)
tures are closer together, there is less time available to 5.7%
CO 12%
prevent thermal runaway using gas detection. In addition,
higher thermal runaway temperatures are better: this means
relatively more energy is needed for the cell to reach that H2
H2 43% CxHy
temperature and go into runaway. CO2 27%
CO2
CO
5.3.2 OFF-GAS FLAMMABILITY
Prior to and during thermal runaway, reactions within the cell CxHy 18%
produce gases which get vented. These gases are considered
the off-gas, which is measured during cell-level UL 9540A
testing to determine flammability and other characteristics. Figure 5-8 Off-gas composition of gas released during thermal
runaway

22
2022 Battery Scorecard

6 2023 AND BEYOND

Driven primarily by the growth in electric vehicle (EV) demand, the lithium batteries we have
today will continue to see advances that benefit the EV sector and, to a lesser degree, the
stationary energy storage sector. In the short- and mid-term, lithium batteries will dominate the
stationary storage market due to their versatility, durability, and general downward trend in cost.
We can also expect improvements across energy density, durability, performance, and safety.
Below we discuss some of the expected improvements.

As background, Li-ion batteries are made of several

components, including an anode, a cathode, electrolyte, ELECTROLYTE
separator, and the cell casing. The anode and cathode are the Advancements are also expected in electrolytes, which
electrochemically active components that store (intercalate) facilitate the flow of ions from anode to cathode and
lithium ions when charging and discharging. Development back again. An example of liquid electrolyte is the acid
efforts have focused on reducing the cost of the cells by in a conventional automotive lead-acid starter battery.
increasing energy density, improving reliability and safety of In lithium batteries, the electrolyte is typically an
the anode and cathode, increasing ability to charge quickly, organic-based compound or solvent that interfaces
and reducing inactive materials. with the cathode and anode and permeates a
separator. While most Li-ion battery electrolytes are
Li-ion battery improvements liquid, polymer electrolytes are also commercially
employed, and solid electrolytes are in development.
The solid electrolyte material is an exciting prospect
CATHODE
because it has a significant safety advantage over
With regards to the cathode, most EV batteries
organic polymer and liquid electrolytes due to its
incorporate some combination of lithium with cobalt,
inflammable nature. Other potential advantages of
magnesium, aluminum, and nickel to form the NMC or
solid-electrolyte batteries include higher energy density,
NCA chemistry, as discussed above. Many of these
longevity, and stability, though these claims need to be
metals, particularly cobalt and lithium, have increased
proven before commercialization.
in cost over the last 12 months due to demand growth,
supply constraints, or both. Lower range EVs and
stationary storage applications also use lithium
ANODE
combined with iron to form the LFP battery chemistry.
Today’s anodes are typically made of graphite, a form of
While LFP is cheaper and generally has a higher
carbon that has a structure which allows lithium ions to
temperature threshold to entering thermal runaway
be reversibly inserted (intercalated) between the carbon
than NMC and NCA, it sacrifices energy density, so
layers. While graphite works quite well as an anode
it takes up more space in the vehicle or project site
material, silicon is being considered as a replacement
for the same amount of storage capacity. It is unclear
since it can hold 10 times more lithium than graphite.
which battery chemistry will prevail as rapid growth and
The problem with silicon is that it swells in size with the
material supply constraints will change the relative costs
insertion of the lithium ions much more than graphite,
of these commodities. Sulfur, as a replacement for the
which can damage the cell. A proposed solution is to
layered oxide class used by NMC and NCA, is also being
use a mixture of graphite and silicon, which has been
considered for battery cathodes because of its very high
seen to reduce swelling at the cost of lesser gains in
energy density, though no commercial products are on
density.
the market.

23
2022 Battery Scorecard

Non-lithium technology Many long-duration storage technologies ‘decouple’ power

A variety of non-lithium technologies are primed to step into and energy, which means they increase duration (MWh) at
the stationary sector, or in the case of pumped hydro, already relatively lower costs without having to increase the power
have a dominant market share. According to the China (MW). This is different than conventional batteries that require
Energy Storage Alliance (CNESA), of the 209.4 GW of proportional increases in power and duration because all DC
electrical energy storage deployed globally at the end of energy storage components are contained within one cell.
2021, pumped hydro had 86.2% of the installed capacity,
followed by Li-ion at 11.0%, and then a variety of other Another advantage of many long-duration technologies is
non-lithium technologies making up the remaining 2.8% their reduced fire safety risk and low degradation potential.
(CNESA Energy Storage Industry White Paper 2022). There are also various disadvantages with long-duration
They can broadly be classified as electrochemical storage energy storage technologies compared to Li-ion batteries,
(other battery chemistries including flow batteries), including:
mechanical storage (compressed/liquid air, gravity storage,
pumped hydro), and thermal storage that converts energy • the round-trip efficiency is limited to approximately
into heat and either directly or indirectly uses that heat to 50%-70%, depending on the technology, compared to
replace electricity. There is also significant focus on Li-ion which often exceeds 90%;
hydrogen, which can be classified as chemical energy • ramp rate for some long-duration energy storage
storage, and how it can play a role in long-duration storage technologies can be limited, meaning they are not well
applications and industrial use in other markets like fertilizer suited for shorter durations under 4 hours;
production and steel processing. DNV is actively working with • they can have moderately to significantly lower energy
various technology providers to independently assess their density compared to Li-ion, meaning they are almost
products, though this year’s Battery Scorecard focuses on always larger and heavier and not well suited for mobile
Li-ion. We hope that future reports will include a wider range applications; and
of technologies. • with a notable exception of hydropower which has been
around for centuries, long-duration energy storage
Many non-lithium energy storage technologies are classified technologies are less proven in many cases, though this is
as ‘long duration’, which the U.S. Department of Energy (DOE) starting to change as investment has been flowing into this
defines as 10+ hours of storage capacity at rated power. sector.
Typically, Li-ion technologies operate for 6 hours or less at
rated power. Long-duration energy storage is an emerging DNV expects non-lithium and long-duration energy storage
focus area that many believe is required to fully decarbonize technologies to gradually expand into niche markets over
the electricity grid. To become viable, the long-duration the next 3-5 years, with increased growth as renewable
storage market must meet the following criteria: penetration increases above 50%. Local and federal
mandates and incentives will also drive the long-duration
• capital costs of long-duration storage equipment must be energy storage market to broader adoption.
significantly cheaper than Li-ion on an average $/kWh
basis (targeting $20/kWh or less), with operation and
maintenance costs and useful life being comparable;
In conclusion, there is a lot to be excited about across
• off-takers (like utilities and commercial & industrial partners)
the energy storage sector in the coming years: strong
must value the service of long-duration storage so that they
growth, technology advancements, policy drivers,
pay project owners enough to cover the cost of operating
improved safety, and increased motivation to support
these assets;
the energy transition away from carbon-based fuels to
• long-duration storage equipment must be as safe or safer
name a few. We at DNV look forward to helping lead
the Li-ion technology; and
the energy transition and would love to hear from you
• performance and safety of long duration storage
on how we can help achieve this together.
equipment must be tested and vetted by third parties to
Want to learn more? Schedule a meeting with our team
build broader confidence in this technology.
to discuss the findings and learn more about what the
Scorecard means for your business.

24
ABOUT DNV
DNV is an independent assurance and risk management provider, operating
in more than 100 countries, with the purpose of safeguarding life, property,
and the environment. As a trusted voice for many of the world’s most
successful organizations, we help seize opportunities and tackle the risks
arising from global transformations. We use our broad experience and
deep expertise to advance safety and sustainable performance, set industry
standards, and inspire and invent solutions.

In the energy industry

We provide assurance to the entire energy value chain through our advisory, monitoring,
verification, and certification services. As the world's leading resource of independent
energy experts and technical advisors, we help industries and governments to navigate the
many complex, interrelated transitions taking place globally and regionally, in the energy
industry. We are committed to realizing the goals of the Paris Agreement, and support our
customers to transition faster to a deeply decarbonized energy system.

Main author: James Daggett

Contributors: Jason Goodhand, Steve Jones, Carrie Kaplan,
Ran Long, Michael Kleinberg, Brian Warshay

Disclaimer DNV
All information is correct to the best of our Utrechtseweg 310-B50 155 Grand Avenue Suite 600
knowledge. Contributions by external authors 6812 AR Arnhem, the Netherlands Oakland CA 94612, USA
do not necessarily reflect the views of the Tel: +31 26 356 9111 Tel: +1 510 891 0446
editors and DNV.
All rights reserved. DNV 2022 Email: contact.energysystems@dnv.com
www.dnv.com