Public Report 7

Lithium-Ion Battery Testing

September 2019
Lithium-Ion Battery Testing ― Public Report 7 II
 About ITP Renewables

ITP Renewables (ITP) is a global leader in energy engineering, consulting and project management, with
expertise spanning the breadth of renewable energy, storage, efficiency, system design and policy.

We work with our clients at the local level to provide a unique combination of experienced energy engineers,
specialist strategic advisors and experts in economics, financial analysis and policy. Our experts have
professional backgrounds in industry, academia and government.

Since opening our Canberra office in 2003 we have expanded into New South Wales, South Australia and New
Zealand.

ITP are proud to be part of the international ITP Energised Group—one of the world’s largest, most respected
and experienced specialist engineering consultancies focussed on renewable energy, energy efficiency and
climate change.

Established in the United Kingdom in 1981, the Group was among the first dedicated renewable energy
consultancies. In addition to the UK it maintains a presence in Spain, Portugal, India, China, Argentina and
Kenya, as well as our ITP offices in Australia and New Zealand.

Globally, the Group employs experts in all aspects of renewable energy, including photovoltaics (PV), solar
thermal, marine, wind, hydro (micro to medium scale), hybridisation and biofuels.

About this report

Supported by a $1.29m grant from the Australian Renewable Energy Agency under its Advancing Renewables
Program, the Lithium-Ion Battery Test Centre program involves performance testing of conventional and emerg-
ing battery technologies. The aim of the testing is to independently verify battery performance (capacity fade and
round-trip efficiency) against manufacturers’ claims.

Six lithium-ion, one conventional lead-acid, and one advanced lead-acid battery packs were installed during
Phase 1 of the trial. The trial was subsequently expanded to include an additional eight lithium-ion packs, a zinc
bromide flow battery, and an aqueous hybrid ion battery bank.

This report describes testing results and general observations or issues encountered thus far with both the
Phase 1 and 2 batteries.

This report, earlier reports, and live test results are published at www.batterytestcentre.com.au.

This Project received funding from ARENA as part of ARENA’s Advancing Renewables Program. The views expressed herein are
not necessarily the views of the Australian Government, and the Australian Government does not accept responsibility for any
information or advice contained within this report.

Lithium-Ion Battery Testing ― Public Report 7 III

 Report Control Record

Document prepared by:

ITP Renewables

Level 1, Suite 1,
19 -23 Moore St, Turner, ACT, 2612, Australia
PO Box 6127, O’Connor, ACT, 2602, Australia

Phone: +61 2 6257 3511

Fax: +61 2 6257 3611
E-mail: info@itpau.com.au
itpau.com.au

Disclaimer:
A person or organisation choosing to use documents prepared by IT Power (Australia) Pty Ltd accepts the
following:

a) Conclusions and figures presented in draft documents are subject to change. IT Power (Australia) Pty Ltd
accepts no responsibility for their use outside of the original report.

b) The document is only to be used for purposes explicitly agreed to by IT Power (Australia) Pty Ltd.

c) All responsibility and risks associated with the use of this report lie with the person or organisation who
chooses to use it.

Lithium-Ion Battery Testing ― Public Report 7 IV

 List of Abbreviations

AC Alternating Current

AIO All-in-one (referring to a battery unit which is combined with a battery inverter and PV inverter)

ARENA Australian Renewable Energy Agency

AUD Australian Dollar

BESS Battery Energy Storage System

BMS Battery Management System

BOS Balance of System

C (number) “C Rate” (charge rate), is a measure of the rate at which the battery is charged/discharged relative
to its nominal capacity. Conversely, it can be thought of as the time over which the entire (nomi-
nal) battery capacity is charged/discharged (ie. a C10 rate indicates a
charge/discharge rate at which a full charge/discharge takes 10 hours. A 2C rate indicates a
charge/discharge rate at which a full charge/discharge takes only 0.5 hours)

CAN (bus) Controller Area Network (a message-based communications protocol allowing microcontrollers
and devices to communicate without a host computer)

DC Direct Current

DOD Depth of Discharge of a battery

ELV Extra Low Voltage

IR Infra-Red (region of the electromagnetic radiation spectrum used in thermal imaging)

ITP IT Power (Australia) Pty Ltd, trading as ITP Renewables

kW Kilowatt, unit of power

kWh Kilowatt-hour, unit of energy (1 kW generated/used for 1 hour)

kWp Kilowatt-peak, unit of power for PV panels tested at STC

LFP Lithium Iron Phosphate (a common li-ion battery chemistry)

Li-ion Lithium-ion (referring to the variety of battery technologies in which lithium ions are
intercalated at the anode/cathode)

LMO Lithium Manganese Oxide (a common li-ion battery chemistry)

LTO Lithium Titanate (a common li-ion battery chemistry)

MODBUS A serial communication protocol for transmitting information between electronic

devices

NMC Nickel Manganese Cobalt (a common li-ion battery chemistry)

NCC National Construction Code

PbA Lead Acid

PMAC Permanent Magnet Alternating Current (a variety of electric motor)

PV Photovoltaic

RE Renewable Energy

SOC State of Charge of a battery

UPS Uninterruptable Power Supply

VRB Vanadium Redox Battery, a type of flow battery

VRLA Valve Regulated Lead Acid

Lithium-Ion Battery Testing ― Public Report 7 V

 Contents

EXECUTIVE SUMMARY 01
1. PROJECT BACKGROUND 02
1.1. Report 1 – September 2016 03
1.2. Report 2 – March 2017 03
1.3. Report 3 – November 2017 04
1.4. Report 4 – March 2018 05
1.5. Report 5 – September 2018 05
1.6. Report 6 – June 2019 05
2. BATTERY OPERATION OVERVIEW 06
3. PHASE 1 UPDATE 07
3.1. Samsung AIO10.8 07
3.2. Sony Fortelion 08
3.3. Tesla Powerwall 1 09
4. PHASE 2 UPDATE 10
4.1. Alpha ESS M48100 10
4.2. BYD B-Box 10
4.3. GNB Lithium 11
4.4. LG Chem RESU HV 12
4.5. Pylontech US2000B 13
4.6. Redflow ZCell 13
4.7. Tesla Powerwall 2 15
5. PERFORMANCE COMPARISON 16
5.1. Phase 1 Capacity Test Results 16
5.2. Phase 2 Capacity Test Results 18
5.3. Round-Trip Efficiency 20
6. MARKET DEVELOPMENT 21
6.1. Cost Trajectory 21
7. LESSONS LEARNED 22
8. KNOWLEDGE SHARING 23
APPENDIX A. TESTING PROCEDURE 26

Lithium-Ion Battery Testing ― Public Report 7 VI

 Executive Summary

ITP Renewables (ITP) is testing the performance

of residential and commercial scale battery
packs in a purpose-built, climate-controlled
enclosure at the Canberra Institute of
Technology. Eight batteries were installed
initially, and a further ten installed in a second
phase. This is the seventh public six-monthly
report.

While some battery packs have experienced

faults and/or failed prematurely, the Sony,
Samsung, Tesla Powerwall 1 (Phase 1), BYD,
Pylontech, and GNB Lithium (Phase 2) battery
packs have generally demonstrated high
reliability, with minimal issues encountered
throughout the testing period.

Linear extrapolation of capacity fade to date

suggests cycle life varies significantly between
products. The Sony, Samsung (Phase 1), and
Pylontech (Phase 2) battery packs continue to
demonstrate good capacity retention over a
large number of cycles. Following replacements,
the current Tesla Powerwall 2 and Redflow ZCell
(Phase 2) are also demonstrating excellent
capacity retention thus far, though the number of
cycles completed to date is low.

Variability in round-trip efficiency is lower, and

has generally been observed between 85-95% for
both the lead-acid and lithium-ion technologies.

With respect to the market at large, price

reductions have stalled in recent months, with
this generally attributed to cell production
constraints and high raw material prices.
Nevertheless, most analysts believe that
manufacturers are substituting away from high
cost inputs, and that the large amount of
production capacity currently under construction
will put downward pressure on prices in the
medium-term. ITP’s opinion is that price
reductions are required for mass-market uptake,
alongside improvements in products, interfaces,
and technical support.

A third phase of battery testing has recently been

announced and will comprise another eight
battery packs, including a lithium-titanate battery
and a sodium-nickel battery. These will replace
eight batteries from Phases 1 and 2 which have
either completed the original testing period or
are no longer cycling for various reasons. Testing
of the remaining Phase 1 and 2 batteries is
continuing.

Lithium-Ion Battery Testing ― Public Report 7 01

 1. PROJECT BACKGROUND

ITP Renewables (ITP) is testing the performance of residential and commercial-scale battery packs in a
purpose-built, climate-controlled enclosure at the Canberra Institute of Technology. The aim of the testing is to
independently verify battery performance (capacity retention and round-trip efficiency) against manufacturers’
claims.

Six lithium-ion, one conventional lead-acid, and one advanced lead-acid battery packs were installed during
Phase 1 of the trial, which commenced in August 2016. The trial was subsequently expanded to include an
additional eight lithium-ion packs, a zinc-bromide flow battery, and an Aquion “saltwater” battery bank. Phase 2
commenced in July 2017.

A Phase 3 has recently been announced and will comprise another eight battery packs, including a
lithium-titanate (LTO) battery and a sodium-nickel battery. The batteries to be installed under Phase 3 are listed
below:

- FIAMM SoNick
- sonnenBatterie
- BYD Battery Box HV
- SolaX Triple Power
- ABB REACT2
- Deep Cycle Systems (DCS) PV 10.0
- Zenaji Aeon
- PowerPlus Energy LiFe Premium

The new batteries will replace eight Phase 1 and Phase 2 batteries to be removed from testing. Batteries still
being cycled from Phase 1 and Phase 2 include:

- Samsung AIO (Phase 1)

- Sony Fortelion (Phase 1)
- Tesla Powerwall 1 (Phase 1)
- BYD B-Box LV (Phase 2)
- GNB Lithium (Phase 2)
- LG Chem RESU HV (Phase 2)
- Pylontech US2000B (Phase 2)
- Redflow ZCell (Phase 2)
- Tesla Powerwall 2 (Phase 2)

Cycling of Phase 3 batteries is scheduled to begin before the release of the next Public Report.

This is the seventh public report outlining the progress and results of the trial thus far. A summary of the six
previous reports is provided below. Complete reports are accessible on the Battery Test Centre website at
www.batterytestcentre.com.au/reports/.

Lithium-Ion Battery Testing ― Public Report 7 02

 1.1. Report 1 ― September 2016

Report 1 was published in September 2016 and outlined the background of the project. The intended audience of
the trial included the general public, research organisations, commercial entities, and government organisations
who are considering investment in battery energy storage.

The report described conventional lead-acid and lithium-ion technologies, the process of battery selection, and
the testing procedure. The implementation process from procurement through installation to commissioning
was also described for the eight Phase 1 batteries listed in Table 1 below.

Table 1. Phase 1 Battery Packs

Product Country of Origin Chemistry Total Installed Capacity (kWh)

CALB CA100 China Lithium Iron Phosphate 10.24

Ecoult UltraFlex USA Lead Acid Carbon 14.8 (C8)

GNB Sonnenschein Germany Lead Acid 14.4 (C100)

Kokam Storaxe Korea Nickel Manganese Cobalt 8.3

LG Chem RESU 1 Korea Nickel Manganese Cobalt 9.6

Samsung AIO Korea Nickel Manganese Cobalt 10.8

Sony Fortelion Japan Lithium Iron Phosphate 9.6

Tesla Powerwall 1 USA Nickel Manganese Cobalt 6.4

At the completion of the first report, battery cycling had been underway for roughly three months. At that early
stage of testing, data did not provide meaningful insight into long-term battery performance. As such, the report
focussed on the lessons learned during the procurement, installation and commissioning phases and set out the
structure in which results would be released in future reports.

1.2. Report 2 ― March 2017

Capacity tests were conducted in each of the six months between September 2016 and February 2017, and the
results were published in Public Report 2.

It was reported that the Kokam Storaxe battery pack had suffered irreversible damage during that time, due to
improper low-voltage protection provided by the built-in Battery Management System (BMS).

It was also reported that the CALB pack required a replacement cell and thereafter was functional, but still
showing evidence of either a weak cell or poor battery management by the external BMS.

Capacity fade was evident for some of the battery packs under test, as expected. However, for others, long-term
trends were not yet discernible owing to the inherent variability in individual capacity test results, attributed to
imprecision in SOC estimation.

Lithium-Ion Battery Testing ― Public Report 7 03

 In terms of round-trip efficiency, despite the limited data, already it could be observed that lithium-ion
out-performs the conventional lead-acid battery pack, despite lead-acid efficiency appearing higher than general
expectations. Refer to the complete report for details.

1.3. Report 3 ― November 2017

Report 3 was published in November 2017. It described the process of procuring and installing the 10 x Phase 2
battery packs listed in Table 2 below, and outlined preliminary testing results and general observations or issues
encountered with the Phase 1 batteries.

Table 2. Phase 2 Battery Packs

Product Country of Origin Chemistry Total Installed Capacity (kWh)

Alpha ESS M48100 China Lithium Iron Phosphate 9.6

Ampetus Super Lithium China Lithium Iron Phosphate 9.0

Aquion Aspen USA Aqueous Hybrid Ion 17.6

BYD B-Box China Lithium Iron Phosphate 10.24

GNB Lithium Germany Nickel Manganese Cobalt 12.7

LG Chem RESU HV Korea Nickel Manganese Cobalt 9.8

Pylontech US2000B China Lithium Iron Phosphate 9.6

Redflow ZCell USA Zinc-Bromide Flow 10.0

SimpliPhi PHI 3.4 USA Lithium Iron Phosphate 10.2

Telsa Powerwall 2 USA Nickel Manganese Cobalt 13.5

In particular, Report 3 described how battery supply and installation issues continued to hamper the progress of
the market as a whole, and that a number of manufacturers had either exited the market or substantially
changing their product offerings. Of further note was that market leaders Tesla and LG Chem had aggressively
cut wholesale pricing, and introduced second generation battery packs.

In terms of Phase 1 pack performance, one Ecoult cell failure was reported and general SOC estimation issues
with the GNB lead-acid battery and Sunny Island inverter were described.

Integration of battery packs with inverters continued to be problematic generally, with the communications
interface being the most common challenge encountered. There was still no standardised approach to
battery-inverter communications and the report described the expectation that installation and commissioning
issues would remain common until communications interface protocols were standardised.

Results from Phase 1 battery pack testing indicated that nascent capacity fade trends were discernible, and that
lithium-ion batteries continued to demonstrate higher efficiency.

Lithium-Ion Battery Testing ― Public Report 7 04

 1.4. Report 4 ― March 2018

Report 4 was published in March 2018. It outlined the preliminary testing results and general issues encountered
with both Phase 1 and Phase 2 batteries. This report provided particular detail on the ongoing commissioning
challenges with the Tesla Powerwall 2 and Aquion battery packs, the replacement of the malfunctioning Redflow
and Ecoult packs, and upgrades to the Ampetus pack.

Ongoing SOC estimation issues for the CALB and GNB lead-acid battery packs were observed, but generally
higher round-trip efficiency for lithium-ion technology over conventional lead-acid and zinc-bromide technologies
continued to be demonstrated.

Capacity test results showed characteristic capacity fade for all Phase 1 battery packs (1,000+ cycles
completed) still in operation. Significant variability between packs was observed, and the potential role of
temperature effects in contributing to these results was discussed. Phase 2 battery packs (500+ cycles
completed) showed similar initial trends and variability in capacity fade.

1.5. Report 5 ― September 2018

With testing of both Phase 1 and 2 batteries well under way by the time Report 5 was published, capacity fade
trends were well-established with significant variation in performance between packs apparent. DC round-trip
efficiency varied less between packs, with average values of 85-95%.

Although several batteries continued to perform well, the report described performance and reliability issues with
some battery packs. In most cases the issues were attributed to inadequate product development and/or a lack
of understanding on the part of local salespeople/technicians in regard to product integration (i.e. with inverters
or control systems).

In particular, the report described the replacement of the Redflow ZCell and SimpliPhi PHI 3.4 packs, ongoing
challenges controlling the Tesla Powerwall 2, the insolvency of Aquion and Ampetus, and some operational
issues with the CALB, LG Chem, EcoUlt and GNB lead-acid Phase 1 battery packs.

1.6. Report 6 ― June 2019

With Phase 1 testing concluding at the end of March 2019, Report 6 included a comprehensive analysis of the
performance of those batteries, as well as an update on Phase 2 batteries. Overall, the Sony (Phase 1) and
Pylontech (Phase 2) battery packs demonstrated excellent capacity retention, and the Sony, Samsung, Tesla
(Phase 1), BYD and Pylontech (Phase 2) battery packs demonstrated high reliability. The Samsung and BYD
battery packs in particular demonstrated consistently high round-trip efficiency.

Round-trip efficiency between 85-95% had been observed for both the lead-acid and lithium-ion technologies,
while linear extrapolation of capacity retention to date suggested that between 2,000-6,000 cycles could be
delivered by properly-functioning lithium-ion battery packs.

The report also discussed the high number of battery packs installed in the Test Centre which had been removed
or replaced prematurely owing to faults. These issues are symptomatic of new technology and a new market,
and are expected to improve over time.

Lithium-Ion Battery Testing ― Public Report 7 05

 2. Battery Operation Overview Figure 1 gives an overview of the issues experienced by battery packs installed in the trial. Note that only
issues inhibiting all cycling are displayed, including commissioning difficulties, failures requiring
replacement, and removal of batteries.

2015 2016 2017 2018 2019

OCT JAN APR JUL OCT JAN APR JUL OCT JAN APR JUL OCT JAN APR JUL

CALB SINGLE FAULTY CELL IDENTIFIED AND REPLACED.

SOC ALGORITHM ERRORS RESULT IN THE

SOC ALGORITHM ERRORS RESULT IN THE BATTERY PACK CYCLING
Ecoult OUTSIDE OF DESIGN PARAMETERS. BATTERY IS REPLACED.
BATTERY PACK CYCLING OUTSIDE OF DESIGN
PARAMETERS. BATTERY TO BE REPLACED.

TESTING
GNB PbA
CONCLUDED
PHASE ONE

Kokam PROCUREMENT, BMS FAILS TO PROTECT BATTERY FROM UNDERVOLTAGE AND IT IS UNABLE TO BE RESTARTED.
INSTALLATION
AND
LG Chem LV COMMISSIONING BATTERY IS REPLACED DUE TO CELL IMBALANCE.

Samsung

Sony

Tesla PW1

BATTERY PACK CYCLING AFFECTED BY OVERTEMPATURE.

Alpha ESS ALPHA REMOVES ITS BATTERIES.

COMMISSIONING
Ampetus DIFFICULTIES.
AMPETUS ENTERS LIQUIDATION.
BATTERY UNABLE TO CYCLE
DUE TO CELL IMBALANCES.
TESTING
AQUION
AQUION RE-ACQUIRED BUT NOT OFFERING CONCLUDED
Aquion FILES FOR BATTERY NOT CYCLING AS UNABLE TO BE COMISSIONED.
SUPPORT FOR EXISTING PRODUCTS.
BANKRUPTCY.

BYD
PHASE TWO

GNB Li-ion PROCUREMENT,

INSTALLATION
AND BMS FAILS TO PROTECT BATTERY FROM UNDERVOLTAGE AND
LG Chem HV COMMISSIONING IT IS UNABLE TO BE RESTARTED. BATTERY IS REPLACED.

Pylontech BATTERY REPLACED DUE

TO ELECTROLYTE LEAK.

BATTERY REPLACED DUE TO BATTERY REPLACED DUE COMMISSIONING OF BATTERY REPLACED DUE
Redflow CONTAMINATED ELECTROLYTE. TO ELECTROLYTE LEAK. SECOND INVERTER. TO ELECTROLYTE LEAK.

SimpliPhi CHANGES TO RECOMMENDED INVERTER SETPOINTS MEAN THAT THE BATTERY PACK HAS TESTING
BEEN CYCLED OUTSIDE OF DESIGN PARAMETERS. SIMPLIPHI REMOVES ITS BATTERIES. CONCLUDED
BATTERY NOT CYCLING AS IT IS UNABLE TO BE BATTERY REPLACED.
Tesla PW2 EXTERNALLY CONTROLLED. CYCLING COMMENCES.

Figure 1: Overview of battery operation

Lithium-Ion Battery Testing ― Public Report 7 06

 3. PHASE 1 UPDATE

This section provides a summary of any developments in the past six months for the remaining Phase 1
batteries, and gives an update on progress overall.

3.1. Samsung AIO10.8

Operational Issues
The Samsung AIO10.8 has completed a high number of cycles. No
faults have been experienced in the past six months or at any time
during testing.

Capacity Fade
The average energy discharged each cycle (Figure 2) can be seen to
have decreased over time, with increasing variance between cycles also
evident. The data suggests a SOH of ~79% after ~2,190 cycles.

12000

10000

8000
Energy Discharged (Wh)

6000

4000

2000

Figure 2. Energy discharged per cycle by the Samsung battery pack

Lithium-Ion Battery Testing ― Public Report 7 07

 3.2 Sony Fortelion

Operational Issues
The Sony pack has completed a high number of cycles. No faults have
been experienced in the past six months or at any time during testing.

Capacity Fade
The average energy discharged each cycle (Figure 3) can be seen to
have generally decreased over time, with greater variance between
cycles also evident. The data suggests a SOH of ~83% after ~2,100
cycles.

9000

8000

7000
Energy Discharged (Wh)

6000

5000

4000

3000

2000

1000

Figure 3. Energy discharged per cycle by the Sony battery pack

Lithium-Ion Battery Testing ― Public Report 7 08

 3.3 Tesla Powerwall

Operational Issues
At the beginning of the trial (Phase 1), Tesla’s Powerwall 1 was only
compatible with a Solar Edge inverter. All other Phase 1 packs, excluding
the Samsung, were compatible with the market-leading SMA Sunny
Island inverter, which the control system had been designed to control.

While ITP was able to control the Solar Edge/Powerwall system via an
online portal, the rate of charge and discharge was not able to be
controlled. Hence, the Powerwall 1 is charging and discharging at its
maximum rate (~2hr full charge/discharge) while other batteries charge
and discharge over ~3hrs. This means the Powerwall has less time to
dissipate heat built up during charge/discharge, which may be causing
higher battery cell temperatures leading to accelerated capacity fade.
Efficiency may also be affected, as the Tesla’s cooling system will be
more heavily loaded. ITP is unable to confirm these hypotheses as the
Tesla system allows for no data access.

Nevertheless, the Tesla Powerwall 1 has proven highly reliable and, in

conjunction with the high allowable DOD, this has allowed the battery
pack to have completed the high number of cycles. No operational
issues have been experienced during testing.

Capacity Fade
The average energy discharged each cycle (Figure 4) can be seen to
have generally decreased over time. The data suggests a SOH of ~64%
after ~2,190 cycles.

6000

5000
Energy Discharged (Wh)

4000

3000

2000

1000

Figure 4. Energy discharged per cycle by the Tesla Powerwall 1 battery pack

Lithium-Ion Battery Testing ― Public Report 7 09

 4. PHASE 2 UPDATE

This section provides a summary of any developments in the past six months for the remaining Phase 2 batteries, and
gives an update on progress overall.

Some battery packs have demonstrated challenges that affect cycling and capacity testing. These issues are described
below.

4.1. Alpha ESS M48100

Operational Issues
During the 2018/19 summer temperature regime, ITP observed that the
Alpha battery pack was constraining the charge and discharge rate
below the rate requested by the test centre’s control system. Alpha
stated that this behaviour is abnormal, and collected the battery pack
for analysis in March 2019. In August 2019 Alpha contacted ITP to say
that the battery had experienced over-temperature alarms, and stated
that they would not be continuing in the battery trial.

No further data is available since publication of the last report and as a

result no analysis is included here.

4.2. BYD B-Box

Operational Issues
ITP has not experienced any operational issues with the BYD battery
pack. However, BYD performed a firmware update on the BMU in June,
and again in August. The BMU was also replaced at that time as it was
unable to accept the firmware update.

Capacity Fade
The energy discharged per cycle is shown in Figure 5. The data
suggests a SOH of ~64% after ~1,740 cycles, with capacity fade
appearing to accelerate and then decelerate. The deceleration may be
the result of the firmware/BMU upgrade.

Lithium-Ion Battery Testing ― Public Report 7 10

 10000

9000

8000
Energy Discharged (Wh)

7000

6000

5000

4000

3000

2000

1000

Figure 5. Energy discharged per cycle by the BYD battery pack

4.3. GNB Lithium

Operational Issues
ITP has not experienced any operational issues with the GNB Lithium
battery pack. When performing diagnostic tests on the battery with
GNB’s proprietary software, a ‘Battery Internal Voltage Too High’ error is
returned. When ITP last contacted GNB, GNB stated that the errors were
regular notifications.

Capacity Fade
The energy discharged per cycle is shown in Figure 6. The data
suggests a SOH of ~65% after ~1,190 cycles. It is notable that some
cycles show capacity has been retained far above the average capacity
delivered each cycle. This suggests the capacity is still available but that
some kind of fault is rendering it unavailable for most of the time.
Nevertheless, GNB have advised that no fault is apparent.

12000

10000
Energy Discharged (Wh)

8000

6000

4000

2000

Figure 6. Energy discharged per cycle by the GNB LFP battery pack

Lithium-Ion Battery Testing ― Public Report 7 11

 4.4. LG Chem RESU HV

Operational Issues
In October 2018, the LG Chem RESU HV battery pack in the trial was
replaced by LG Chem, as the previous unit experienced swelling of the
battery cells and undervoltage after a period of disconnection. LG Chem
has advised that the deep self-discharge is due to the internal DC-DC
converter staying on and consuming energy from the battery. Since
publication of the last report, LG Chem has developed an improved
battery model which prevents deep self-discharge with a switch
between the DC-DC converter and the battery cells. LG Chem offered a
replacement unit to ITP; however, ITP has chosen to retain its current
battery in order to continue testing. LG Chem has advised that it will be
replacing units which experience undervoltage problems.

Capacity Fade
The energy discharged per cycle is shown in Figure 7. The data
suggests a SOH of ~90% after ~620 cycles.

8000

7000

6000

5000
Energy Discharged (Wh)

4000

3000

2000

1000

Figure 7. Energy discharged per cycle by the LG Chem RESU HV battery pack

Lithium-Ion Battery Testing ― Public Report 7 12

 4.5. Pylontech US2000B

Operational Issues
ITP has not experienced any operational issues with the Pylontech
battery pack.

Capacity Fade
The energy discharged per cycle is shown in Figure 8. The data
suggests a SOH of ~86% after ~1,470 cycles.

10000

9000

8000

7000
Energy Discharged (Wh)

6000

5000

4000

3000

2000

1000

Figure 8. Energy discharged per cycle by the Pylontech battery pack

4.6. Redflow ZCell

Operational Issues
The Redflow battery suffered an electrolyte leak and was replaced in
February 2019. This was the fourth time the Redflow battery has been
replaced in this trial, and the third time it has been replaced due to an
electrolyte leak. The first replacement was due to contaminated electrolyte.

Redflow attributed the leak to a step in their manufacturing process in

which the electrolyte tank was washed with a particular soap after
manufacture, causing brittleness in the plastic and therefore increased risk
of cracks. This apparently only affected a specific batch of products.
The previous leaks were attributed to micro-cracking of the electrolyte tank
that occurred during road transport. The problem identified was that the
electrolyte trays were not sufficiently supported on the sides to withstand
the weight of the electrolyte. Redflow state that they have since modified
their transport techniques and believe this problem will be avoided in the
future.

Lithium-Ion Battery Testing ― Public Report 7 13

 Since then, the Redflow battery has not experienced any problems, and has been cycling well. Redflow staff visited the test
centre in September 2019 to check the pH of the electrolyte, and reported that it was satisfactory.

The Redflow battery also operates on a slightly different cycling regime to other batteries in the trial. Due to battery charge
rate limits, as well as the requirement for regular maintenance cycles during which normal operation is paused, the Redflow
only completes two full cycles per day (instead of three).

The purpose of the maintenance is to remove all zinc from the electrode stack so the next charge cycle starts with a “clean
slate”. The maintenance cycle requires the battery be fully discharged before the maintenance can occur, and in the trial
set-up this occurs at the end of each day (after two complete cycles).

Capacity Fade
The energy discharged per cycle is shown in Figure 9. The data suggests a SOH of 100% after ~370 cycles.

14000

12000

10000
Energy Discharged (Wh)

8000

6000

4000

2000

Figure 9. Energy discharged per cycle by the Redflow battery pack

Lithium-Ion Battery Testing ― Public Report 7 14

 4.7. Tesla Powerwall 2

Operational Issues
In September 2018, the Tesla Powerwall 2 identified a ‘welded relay’
fault. Tesla suggested that this may have been related to the burnt-out
terminal block discovered following installation, although this was not
confirmed and it is unclear what caused the fault. Both the Powerwall 2
and associated Gateway were subsequently replaced by Tesla. Cycling
of the replacement Powerwall 2 commenced in late November.

ITP still have no direct control over the battery (as Tesla do not allow
this level of control of their products), but rely on Tesla to implement the
cycling schedule. This has generally worked well; however, in June 2019
ITP noted that the discharge power appeared to be fluctuating, and that
as a result the battery wasn’t always reaching the minimum SOC every
cycle. Tesla has stated that the Powerwall 2 inverter is turning off due to
overvoltage (as required by Australian Standards). This may be a result
of the Battery Test Centre electrical connection being re-located due to
electrical works at the CIT. The issue appears to have abated in recent
months.

User-friendly monitoring of the Tesla Powerwall 2 is only possible via

mobile app. Data is available from the Tesla Powerwall 2’s local web
interface. Although Tesla has not published local API documentation,
community groups of have published a tutorial on how to take data from
the battery online. The data used by ITP in monitoring and analysis is
1

obtained from this API.

Capacity Fade
The energy discharged per cycle is shown in Figure 10. The data
suggests a SOH of ~96% after ~640 cycles.

14000

12000

10000
Energy Discharged (Wh)

8000

6000

4000

2000

0
Figure 10. Energy discharged per cycle by the Tesla Powerwall 2 battery pack

1
mikesgear.com/2017/12/07/monitoring-teslas-powerwall2-on-pvoutput-org/

Lithium-Ion Battery Testing ― Public Report 7 15

 5. PERFORMANCE COMPARISON

Testing the capacity of a battery cell involves discharging the cell between an upper and lower voltage limit at a fixed
current, at a given ambient temperature. Because ITP is conducting pack-level testing, the upper and lower voltage limits
are not accessible, and hence the maximum and minimum SOC must be used as a proxy. The result is that the precision of
a single capacity test depends significantly on the SOC estimation, conducted either by the battery inverter/charger or the
in-built BMS.

Throughout the trial, ITP has observed erratic SOC estimation resulting in significant variability in the energy discharged
each cycle. As such, this report provides data and analysis based on both the energy discharged during the monthly
capacity tests (below), as well as on the energy discharged each “cycle” over the course of the trial (see Sections 3 and 4
above, where a cycle is defined as a continuous discharge exceeding 40 minutes in length). Both data sets should be
considered before drawing conclusions.

5.1. Phase 1 Capacity Test Results

Figure 11 shows the estimated state of health (SOH) against cycles completed for each Phase 1 battery pack still cycling.
SOH is estimated by dividing the energy delivered at each capacity test by the energy delivered in the first capacity test.

Tesla Powerwall 1 Samsung AIO Sony Fortelion

110%

100%

90%
Estimated State of Health

80%

70%

60%

50%
0 250 500 750 1000 1250 1500 1750 2000 2250 2500

Cycle Count (Energy Discharged/Nameplate Capacity)

Figure 11. Capacity fade of Phase 1 battery packs based on monthly capacity tests

It should be noted that Figure 11 includes lines-of-best-fit that are determined by simple linear regression. While a linear
regression appears to provide a good fit to the capacity test data collected thus far, extrapolating linearly into the future may
not be appropriate.

Lithium-Ion Battery Testing ― Public Report 7 16

 Samsung AIO10.8

The most recent capacity test suggests a SOH of 82%, broadly in

agreement with the 79% SOH estimated from cycle data (described in
Section 3.1).

Based on the linear regression between estimated SOH and cycles

completed (Figure 11), the Samsung AIO pack is on track for 60% SOH
at ~3,770 cycles. As above, however, the cycle data suggests some
non-linearity which may invalidate this extrapolation.

Sony Fortelion

The most recent capacity test suggests a SOH of 85%, broadly in

agreement with the 83% SOH estimated from cycle data (described in
Section 3.2).

Based on a linear regression between estimated SOH and cycles

completed (Figure 11), the Sony Fortelion pack is on track for 60% SOH
at ~5,640 cycles. As above, however, a linear extrapolation may not be
appropriate.

Tesla Powerwall 1

The most recent capacity test suggests a SOH of 61%, broadly in

agreement with the 64% SOH estimated from cycle data (described in
Section 3.3).

Based on a linear regression between estimated SOH and cycle count

(Figure 11), the Tesla Powerwall 1 is on track for 60% SOH at ~2,310
cycles. As above, however, the cycle data suggests some non-linearity
which may invalidate this extrapolation.

Lithium-Ion Battery Testing ― Public Report 7 17

 5.2. Phase 2 Capacity Test Results

Figure 12 shows the estimated state of health (SOH) against cycles completed for each Phase 2 battery pack
still cycling. SOH is estimated by dividing the energy delivered at each capacity test by the energy delivered in the
first capacity test.

Pylontech LG Chem RESU HV BYD B-Box LV GNB Li-ion

110%

100%

90%
Estimated State of Health

80%

70%

60%

50%

0 200 400 600 800 1000 1200 1400 1600 1800 2000

Cycle Count (Energy Discharged/Nameplate Capacity)

Figure 12. Capacity fade of Phase 2 battery packs based on monthly capacity tests

It should be noted that Figure 12 includes lines-of-best-fit that are determined by simple linear regression. While
a linear regression appears to provide good fit to some of the capacity test data collected thus far, extrapolating
linearly into the future may not be appropriate.

BYD B-Box

The most recent capacity test suggests a SOH of 64%, in agreement

with the 64% SOH estimated from cycle data (described in Section 4.2).
Based on the linear regression between estimated SOH and cycles
completed (Figure 12), the BYD B-Box is on track for 60% SOH at ~1,960
cycles. As above, however, the data suggests some non-linearity which
may invalidate this extrapolation.

Lithium-Ion Battery Testing ― Public Report 7 18

 GNB Lithium

The most recent capacity test suggests a SOH of 65%, in agreement

with the 65% SOH estimated from cycle data (described in Section 4.3).

Based on the linear regression between estimated SOH and cycles

completed (Figure 12), the GNB Lithium is on track for 60% SOH at
~1,360 cycles. As above, however, the data suggests some non-linearity
which may invalidate this extrapolation.

It should be noted that the previous report assumed a higher nameplate

capacity when determining the cycle count at each capacity test. As a
result, a lower rate of capacity fade appears in this report. The previous
assumption was that the nameplate capacity was as per the
brochures/manuals provided to ITP by the supplier at the time of
purchase. In this report, the nameplate capacity has been assumed as
per the actual nameplate on the battery, which aligns better with the
specifications provided by diagnostic reports produced by the in-built
BMS.

LG Chem RESU HV

The most recent capacity test suggests a SOH of 91%, broadly in

agreement with the 90% SOH estimated from cycle data (described in
Section 4.4).

Based on the linear regression between estimated SOH and cycles

completed (Figure 12), the LG Chem RESU HV is on track for 60% SOH
at ~3,080 cycles. As above, however, a linear extrapolation may not be
appropriate.

Pylontech US2000B

The most recent capacity test suggests a SOH of 89%, broadly in

agreement with the 86% SOH estimated from cycle data (described in
Section 4.5).

Based on the linear regression between estimated SOH and cycles

completed (Figure 12), the Pylontech US2000B is on track for 60% SOH
at ~4,460 cycles. As above, however, a linear extrapolation may not be
appropriate.

Lithium-Ion Battery Testing ― Public Report 7 19

 Redflow ZCell

The Redflow ZCell is controlled via the ZCell portal, where it follows a
daily cycling regime. The portal does not currently allow for monthly
scheduled changes to implement the capacity test regime. Though few
cycles have been completed to date, from the cycling data shown in
Figure 9, no capacity fade is apparent.

Tesla Powerwall 2

The Tesla Powerwall 2 cycling regime is implemented by Tesla, based

on requests from ITP. However, capacity tests for the Tesla Powerwall 2
have suffered from the overvoltage issue mentioned in Section 4.7,
resulting in intermittent cycling. From the cycling data shown in Figure
10, only a small amount of capacity fade is evident thus far.

5.3. Round-Trip Efficiency

The lifetime round-trip efficiency results are shown for each battery in Figure 13. Note that the result shown for
the Tesla PW2 in orange is the AC round-trip efficiency. DC values are not available for the PW2, but can be
assumed to be higher.

100%

90%
Estimated State of Health

80%

70%

60%

50%
BYD B-Box GNB LFP LG Chem HV Pylontech Redflow Samsung Sony Tesla PW1 Tesla PW2

Figure 13. Lifetime round-trip efficiency for each battery pack

It is apparent that the lithium-ion battery packs outperform the Redflow zinc-bromide flow battery pack.

Lithium-Ion Battery Testing ― Public Report 7 20

 6. MARKET DEVELOPMENT

6.1. Cost Trajectory

Since the beginning of the project, the cost of residential and commercial scale lithium-ion battery packs has
fallen significantly. Further, throughout that period, many manufacturers have significantly altered their product
offering, and several have exited the market or become insolvent. In recent periods, cost progress has slowed,
attributed to capacity constraints at the manufacturing level and increasing raw material costs (cobalt, in
particular).

At the same time, the established conventional lead-acid market has been stable, with product prices following
currency and lead price fluctuations.

These trends have continued since publication of the last Public Report.

Wholesale Lithium Battery Prices

$2500

LFP 1
LFP 2
LFP 3
$2000 LFP 4
LFP 5
LFP 6
LFP 7
NMC 1
$1500
NMC 2
NMC 3
NMC 4
NMC 5
$1000 NMC 6
NMC 7
Linear (LFP 3)
Linear (LFP 5)
Linear (NMC 1)
$500 Linear (NMC 3)
Linear (NMC 4)
Linear (NMC 5)

$0
Dec-14 Jul-15 Jan-16 Aug-16 Mar-17 Sep-17 Apr-18 Oct-18 May-19 Dec-19

Figure 14: Wholesale prices for lithium-ion battery products installed in the Battery Test Centre

Significant lithium-ion production capacity is coming online in the medium term, and manufacturers are
increasingly substituting cobalt out of their cells. The effect should be falling lithium-ion costs in the
medium-term.

Lithium-Ion Battery Testing ― Public Report 7 21

 7. LESSONS LEARNED

Having been in operation for almost three years now, the Battery Test Centre project has revealed a number of
valuable lessons. The lessons learned relate not only to the performance of the batteries throughout the trial
(analysed in Sections 3 and 4), but also to the performance of suppliers in delivering products and providing
technical support during commissioning and operation. These lessons were described in Report 6, which was
a major report coinciding with the conclusion of Phase 1 testing. While all of those lessons are still pertinent,
the following observations have also been made since the last Public Report:

- The market appears to be moving towards either integrated battery and inverter products, or battery packs
that are only compatible with inverters from the same manufacturer. ITP experienced many integration
issues between batteries and inverters during the commissioning of Phases 1 and 2. A single integrated
product, or compatibility only between products from the same manufacturer, removes the requirement for
manufacturers to undertake R&D, testing, and maintenance with external partners. It also provides a single
point of accountability for users who experience system problems.

- More high-voltage battery inverters and battery packs are now available. High-voltage battery products are
generally simpler to install, due to smaller cables being required. Higher-voltage inverters are generally more
efficient and have higher power density, meaning cheaper equipment and easier/cheaper installation.

Lithium-Ion Battery Testing ― Public Report 7 22

 8. KNOWLEDGE SHARING

An important part of the battery testing project has been to maximise the demonstration value of the trial by:
- Sharing the knowledge with the largest possible audience
- Publishing data in a way that is highly accessible and user friendly
- Adding value to the raw data through expert analysis and commentary

The Knowledge Sharing seeks to publicise data and analysis generated by the battery testing in order to help
overcome the barriers impeding the up-take of battery storage technology. In particular, it seeks to overcome
the barrier that there are no known published studies of side-by-side battery comparisons which test
manufacturers’ claims about battery performance. This lack of independent verification contributes to investor
uncertainty.

The intended users of the information generated by the project include:

- Future energy project developers, including technology providers and financiers, who will be examining the
investment case of a range of energy storage options.

- Energy analysts involved in projecting future renewable energy costs and uptake rates.
- Electricity industry stakeholders including generators, TNSPs, DNSPs, and regulators.

The Battery Test Centre website was established as the key mechanism for this Knowledge Sharing. The
2

website includes background on the project, live tracking of battery status, and a virtual reality component that
replicates the battery test facility. To date the site has had over 167,150 page views with an average of 2:03
minutes spent per page overall and 3:53 minutes spent on the reports page.

2
batterytestcentre.com.au

Lithium-Ion Battery Testing ― Public Report 7 23

 1 27,423 Figure 15: Number of sessions by country

The data from the website shows that the key audience is Australia, with Australian IP addresses accounting for
41,267 sessions (50%). A session is logged as a single viewer who may view multiple pages within a restricted
period (periods are normally reset after 30 minutes of inactivity). Australia is followed by 9,501 sessions from the
United States, 2,755 from the United Kingdom and Germany not far behind on 2,727. It is interesting to note,
however, that the content has been accessed from right across the globe.

1200

1000

800

600

400

200

0
OCT-16 JAN-17 APR-17 JUL-17 OCT-17 JAN-18 APR-18 JUL-18 OCT-18 JAN-19 APR-19 JUL-19

Figure 16: Weekly active users

Figure 16 above shows the number of weekly active users that have accessed the website and there is a clear
rise between the Phase 1 figures at around 250 weekly users, to the launch of Phase 2 in August of 2017 when
the weekly averages nearly doubled to around 500 active weekly users. The peaks coincided with media articles
that were distributed on those dates.

Lithium-Ion Battery Testing ― Public Report 7 24

 Since then the number of users has been on a gradual upwards trajectory, with an increase noted after the
release of Report 6 and associated media articles in June 2019. The number of weekly users currently hovers
around 500.

There is a good spread of views across the website, particularly the technology and results pages; the top five
most viewed pages after the homepage (18%) are the results page (12%), LG Chem RESU (9%), the reports page
(9%), Pylontech US2000B (6%) and the background page on lithium-ion technology (4%).

Background - Lithium-ion
4%
Pylontech US2000B
6%

Reports
9%

LG Chem RESU
9%
Other
42%

Results
12%

Homepage
18%

Figure 17: Breakdown of the 167,150 page views

Lithium-Ion Battery Testing ― Public Report 7 25

 APPENDIX A. TESTING PROCEDURE

The key objective of the testing is to measure the batteries’ decrease in storage capacity over time and with
energy throughput. As the batteries are cycled they lose the ability to store as much energy as when they are
new.

To investigate this capacity fade, the lithium-ion batteries are being discharged to a state of charge (SOC)
between 5% and 20% (depending on the allowable limits of the BMS), while the lead-acid batteries are being
discharged to a 50% SOC (i.e. 50% of the rated capacity used). The advanced lead battery is being be cycled
between 30% and 80% SOC. These operating ranges are in line with manufacturers’ recommendations for each
technology.

Each battery pack is charged over several hours (mimicking daytime charging from the PV), followed by a short
rest period, then discharged over a few hours (mimicking the late afternoon, early evening period) followed by
another short rest period. In total, there are three charge/discharge cycles per day.

Temperature Profile

The ITP lithium-ion battery trial aims to test batteries in ‘typical’ Australian conditions. It is expected that most
residential or small commercial battery systems will be sheltered from rain and direct sunlight, but still be
exposed to outdoor temperatures; therefore, the ambient temperature in the battery testing room is varied on a
daily basis, and varies throughout the year. The high and low temperatures are given in Table 1.

ITP implements ‘summer’ and ‘winter’ temperature regimes for the three daily charge/discharge cycles. In the
summer months the batteries undergo two cycles at the monthly high temperature and the third at the monthly
low temperature, and in the winter months the batteries undergo two cycles at the monthly low temperature and
the third at the monthly high temperature.

Table 3: Daily high and low ambient temperatures throughout the year

JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

Low 22 20 18 16 14 12 10 12 14 16 18 20

High 36 34 32 30 28 26 24 26 28 30 32 34

Regime S S S S W W W W W W S S

Lithium-Ion Battery Testing ― Public Report 7 26

 40

35

30
Ambient Temp (°C)

25

20 Hot Cycle

Cold Cycle
15

10

0
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC

Figure 18: Daily hot and cold cycle temperatures throughout the year

Given the focus on energy efficiency and low energy consumption at the CIT Sustainable Skills Training Hub, the
timing of the high and low temperature cycles is matched with the variations of outdoor temperatures, to allow
transitions between high and low temperature set-points to be assisted by outdoor air. The schedule of charge
and discharge cycles is show in Figures 2 and 3.

Summer Temperature Regime

100

80
State of charge (%)

60

40

20

0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
:0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0 :0
01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 00 01

Figure 19: Summer temperature regime and charge regime

Lithium-Ion Battery Testing ― Public Report 7 27

 Winter Temperature Regime

100

80
State of charge (%)

60

40

20

0
0 0 0 0 0 0 0 0 0
:0 :0 3:0 4:0 5:0 6:0 7:0 8:0 9:0 0:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 0:00 1:00 2:00 3:00 0:00 1:00
01 02 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 2 2 2 2 0 0

Figure 20: Winter temperature regime and charge regime

On the last day of each month, the batteries undergo a charge/discharge cycle at 25 °C as this is the reference
temperature at which most manufacturers provide the capacity of their batteries. From this, an up-to-date
capacity of the battery can be obtained and compared to previous results to track capacity fade. Although the
duration of a month varies between 28 and 31 days, ITP does not expect this to make a statistically relevant
difference to the results.

Lithium-Ion Battery Testing ― Public Report 7 28

Lithium-Ion Battery Testing ― Public Report 7 29
ITP RENEWABLES

Office: Level 1, 19 Moore St,

Turner, ACT 2612

Postal: PO Box 6127, O’Connor,

ACT 2602, Australia

Email: info@itpau.com.au
Phone: +61 (0) 2 6257 3511
Fax: +61 (0) 2 6257 3611

ITPAU.COM.AU

Follow Us: