DNV GL © SAFER, SMARTER, GREENER DNV GL ©

Sensor Enhanced and Model Validated Life Extension of Li-Ion Batteries for Energy Storage

1

DNV GL, NexTech Materials, Beckett Energy Systems

DNV GL ©

Value Propositions (derived from list of Battery Challenges)

(Safety) Prognostic approach for battery failure

Controlling battery packs to extract maximal capacity

Enables greater safety for fast charging

Assurance against “perceived” liabilities

General Safety and prevention – multiple inquiries

<5% system cost but > 5% value add?

Integration flexibility

Value in harsh climates

Cost vs. complexity

Non xEV applications (grid storage…)

Failure modes

Modularity of system

Battery and BMS Failure Modes

Expanding pack lifetime or operational limits

2

DNV GL ©

Safety Prognostic Approach to Battery Failure

3

240 245 250 255 260 265 2700

50100

150200250

300350400

Tem

pera

ture

(d

eg

C)

240 245 250 255 260 265 2700

2040

6080100

120140160

Sen

sor

Resp

onse

(kO

hm

)

Battery Surface Temp

Air Temp

240 245 250 255 260 265 2700

3

6

9

Volt

age

(V

)

Time (minutes)

240 245 250 255 260 265 2700

50

100

150

Sen

sor

Resp

onse

(kO

hm

)

Cell Voltage

Sensor Signal

Thermal Runaway Indicated

by Temperature Spike

Sensor Indication of Off-Gassing

Voltage

Indication of

Failure

Overcharge Test

DNV GL ©

Industry-Wide Problem

4 4

Automated Halogen

Extinguisher

DNV GL ©

Safety & Reliability

5

Safety and Reliability

Prognositics

Performance

Off gas

monitoring

DNV GL ©

Offgas Monitoring Enables Appropriate Use of Automated Fire Extinguishing

6

2 minutes

ahead of

voltage

7-8 minutes

total

7.4 minutes

ahead of

temperature

DNV GL ©

Failure Mode Identification

7

Voltage

2 Minutes

Tem

pera

ture

7 Minutes

Tota

l Early

Warn

ing

Regim

e

7 Minutes – Days

Safety Prognostics

Nail puncture

test – coincident

with V and T

DNV GL ©

Integration Flexibility, Cost vs. Complexity, and Modularity

1-3 sensors within a CES unit

8

• Pack level monitoring

• Benign offgas events are

detectable within minutes

• Catastrophic early warning

detectable < 1 min

DNV GL ©

Control Strategies to Convert to a Binary Signal

9

200 205 210 215 220 225 2300

500

Tem

pera

ture

(d

eg

C)

200 205 210 215 220 225 2300

200

Sen

sor

Resp

onse

(kO

hm

)

Battery Surface Temp

Air Temp

200 205 210 215 220 225 2300

5

10

15

20

Volt

age

(V

)

Time (minutes)

200 205 210 215 220 225 2300

100

200

Sen

sor

Resp

onse

(kO

hm

)

n=100, 2x std dev

n=100, 3x std dev

n=2000, 2x std dev

n=2000, 3x std dev

Voltage, 2x std dev

Voltage, 3x std dev

DNV GL ©

Detection Rates based on Diffusion of Offgas in CES Unit

10

DNV GL ©

Extraction of Maximal Capacity and Extension of Operational Limits

Use sensor as safety assurance for:

11

Increased energy services

Extending cell voltage limits

Expanding C-rate (power) capability

Extending Life or Cumulative Throughput

More value

from the

same system

More

Energy

More

Power

More

Throughput

DNV GL ©

What is the best path for cost reduction?

Monitoring and more aggressive

cycling (A)

Systems that have monitoring

and life extension (L)

Monitoring, life extension, and

more aggressive treatment (LA)

Life extension and downsizing

(LD)

Life extension and bypassable

modules (LB)

Life extension, downsizing, and

bypassable modules (LDB)

Life extension, aggressive

cycling, downsizing, and

bypassable modules (LDBA)

12

0

5

10

15

20

25

30

35

2010 2015 2020 2025 2030 2035 2040 2045

Lifetime Throughput

Monitored System Capacity (kWh)

ConventionalSystem Capacity (kWh)

Monitored 25th Capacity (kWh)

Monitored 75th Capacity (kWh)

Conventional 25th Capacity (kWh)

Conventional 75th Capacity (kWh)

$0.00$0.00

$0.05

$0.01

$0.01

$0.06

$0.00

$0.01

$0.02

$0.03

$0.04

$0.05

$0.06

$0.07

$0.08

Minimum $/kWh,Discharge Only

(Islanding)

Minimum $/kWh PaidCharge and Discharge

(Regulation)

Minimum $/kWh, PaidDischarge, Cost Charge

(Load Shifting)

$/k

Wh

CAPEX/Throughput

Sensor Max

Sensor Mean

Sensor Min

Conventional Max

Conventional Mean

Conventional Min

78%

-63%

-63%

-14%

Monitored System Life Output Improvement

"Islanding" Mean Cost Difference

"Regulation" Mean Cost Difference

"Load Shifting" Mean Cost Difference

Monitored System vs. Conventional System

$38,653

$49,500

Monitored System CAPEX Conventional System CAPEX

20

30

Monitored System Size(kWh)

Conventional System Size(kWh)

3,535,286

1,991,544

Monitored Discharge kWh Conventional Discharge kWh

Total Discharged kWh

DNV GL ©

Assurance against perceived and real liabilities

Perception of decreased safety in longer life

Real consideration for monitoring in enclosed spaces

Real consideration for monitoring in populated areas or the built environment

13

𝐴𝑣𝑜𝑖𝑑𝑒𝑑 𝐶𝑜𝑠𝑡 = 𝑝𝑑 ∗ 𝑝𝑐 ∗ 𝐶𝑐

𝐴𝑑𝑑𝑖𝑡𝑖𝑜𝑛𝑎𝑙 𝑅𝑒𝑣𝑒𝑛𝑢𝑒 = 𝐿 ∗ 𝑅𝑘𝑊ℎ ∗ 𝐸𝑑 ∗ 𝐿𝑒𝑥 (2)

Where pc is the probability of a catastrophe, pd is the probability of detection, and Cc is the costs of that catastrophe.

If the revenue potential per kWh is RkWh, and there is the ability to deliver energy at the rate of Ed (kWh/mo), the expected lifetime of a non-sensor equipped system is L (in months), and the potential life extension percentage is Lex,

Catastrophe

avoidance

Life Extension

DNV GL ©

Additional Cost Opportunity

14

𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑦 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 𝐶𝑜𝑠𝑡 𝑅𝑒𝑑𝑢𝑐𝑡𝑖𝑜𝑛

= 𝐵𝑎𝑠𝑒𝑙𝑖𝑛𝑒 𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑦 − 𝐷𝑜𝑤𝑛𝑠𝑖𝑧𝑒𝑑 𝐶𝑎𝑝𝑎𝑐𝑖𝑡𝑦 ∗$𝑐𝑎𝑝𝑎𝑐𝑖𝑡𝑦

𝑘𝑊ℎ− 𝑆𝑒𝑛𝑠𝑜𝑟 𝑆𝑦𝑠𝑡𝑒𝑚 𝐶𝑜𝑠𝑡

(4)

𝐶𝑜𝑜𝑙𝑖𝑛𝑔 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 𝐶𝑜𝑠𝑡 𝑅𝑒𝑑𝑢𝑐𝑡𝑖𝑜𝑛

= 𝐵𝑎𝑠𝑒𝑙𝑖𝑛𝑒 𝐶𝑜𝑜𝑙𝑖𝑛𝑔 − 𝐷𝑜𝑤𝑛𝑠𝑖𝑧𝑒𝑑 𝐶𝑜𝑜𝑙𝑖𝑛𝑔 ∗$𝑐𝑜𝑜𝑙𝑖𝑛𝑔

𝑘𝑊ℎ− 𝑆𝑒𝑛𝑠𝑜𝑟 𝑆𝑦𝑠𝑡𝑒𝑚 𝐶𝑜𝑠𝑡

(5)

Reducing

cooling load

Reducing

overcapacity

DNV GL ©

Non xEV Applications

15

Offshore oil & gas

- Dynamic Positioning Shuttle Tanker =18 MW for 30 minutes of backup

power for maneuvering during a blackout. Vessel is 300 m long.

- Drilling applications – 5-6 MW transients in <1 minute

- Harbor and offshore tugs: 7-8 MW electric only pulling

DNV GL ©

Harsh Environments < -- > Extended Performance

16

Higher temperature

Cooling load and burden

Physical, ingress, humidity, or mechanical hazards

DNV GL ©

Battery and BMS failure modes

17

Integrated with BMS Independent

of BMS

• Automated

Class D

Extinguisher

• Redundancy

• Maintenance

Alarms

• Distinguish

cell anomaly

from

catastrophic

failure

• Expanded

Operation

• “Pull back”

from

extremes

• Redundancy

in Shut Down

• Bypassable

Modules

DNV GL ©

SAFER, SMARTER, GREENER

www.dnvgl.com

18