Study of DC Arc Flash Phenomenon From 1000 V Lithium Ion Battery Systems

Overview

The research question is how much heat, light and sound a DC arc flash from lithium-ion batteries at application-relevant voltages produces, and how well existing estimation methods match it. EPRI and UTA performed 91 arc flash experiments on LFP (lithium iron phosphate) and LTO (lithium titanate) sources at up to about 1000 VDC, measuring incident energy, arc duration, sound pressure and light intensity. The authors present this as the first comprehensive experimental analysis of DC arc flash from 1000 V lithium-ion systems.

Measured energies were compared with the maximum-power (Doan) method, Ammerman's arc-power model and a fraction-of-bolted-fault-power approach. LFP modules at 540 V produced up to 6.12 cal/cm² from a 2.39 s arc at a 0.25 in gap; at 908 V, LFP gave up to 4.48 cal/cm² (0.5 in gap, manually cut off); LTO at 730 V gave 2.16 cal/cm² for a 0.88 s arc. Arcs below 270 V did not exceed 200 ms, but at 536 V arcs lasted up to 2.94 s (4P) and a 919 V test lasted 4.32 s before manual disconnection.

A worked example of the conventional estimate gave 19.11 cal/cm² with no OCPD and 6.70 cal/cm² with OCPD before applying an enclosure multiplier of up to 3, values the authors consider unrealistic for such battery systems. The maximum-power model overestimated a 0.89 s LTO arc by more than 4 times (up to 8.99 cal/cm², implying category 2 PPE), whereas Ammerman's model agreed well with measured values for both chemistries.

Limits: Ammerman's model assumes the arc distributes its energy equally in all directions in open air, and the efficiency with which an enclosure would absorb and reflect heat is unknown. For short arcs the thermocouples could not measure meaningful energy. The authors stress that long, self-sustaining arcs at higher voltages and non-thermal hazards (sound, light) still matter; measured pressures were not at levels causing lung or ear damage, but hearing and vision protection is recommended.

Key points

1. 91 arc-flash experiments were performed on LFP and LTO lithium-ion battery sources at voltages up to roughly 1000 VDC.
2. LFP modules at 540 V produced up to 6.12 cal/cm² from a 2.39 s arc at a 0.25 in gap; at 908 V, LFP gave up to 4.48 cal/cm² (0.5 in gap, manually cut off); LTO at 730 V gave 2.16 cal/cm² for a 0.88 s arc.
3. A worked conventional estimate gave 19.11 cal/cm² with no OCPD and 6.70 cal/cm² with OCPD before applying an enclosure multiplier of up to 3, values the authors consider unrealistic.
4. Arcs below 270 V did not exceed 200 ms, but at 536 V arcs lasted up to 2.94 s (4P) and a 919 V test lasted 4.32 s before manual disconnection.
5. The maximum-power model overestimated a 0.89 s LTO arc by more than 4 times (up to 8.99 cal/cm², category 2 PPE); Ammerman's model agreed well with measured values for both chemistries.
6. The authors recommend hearing and vision protection because of measured sound and light intensity, noting that measured pressures were not at levels causing lung or ear damage.

Using it in design

• In arc flash estimation for high-voltage battery rooms, do not use the maximum-power method alone with an enclosure multiplier; cross-check unrealistic results with measurement-based models such as Ammerman's.
• Account for arc duration lengthening as voltage rises (536 V and above) when designing battery string protection and disconnection strategy (OCPD, remote disconnect).
• Add hearing and vision protection to the PPE list for battery-area work; arc sound and light intensity are hazards beyond thermal calculation.

Who should read it

Electrical safety engineers carrying out arc flash studies for lithium-ion battery rooms, BESS and UPS batteries.

Document contents

1. I. Introduction (A. Example case study, B. Literature review)
2. II. Test Environment (A. Battery model and equivalent circuit, B. Arc in a box, C. Sensor board)
3. III. LFP Arc Flash Experiments (A. IE, B. Arc duration, C. Sound pressure, D. Light intensity)
4. IV. LTO Arc Flash Experiments
5. V. Arc Power Analysis
6. VI. IE Model Comparison
7. VII. Conclusion

Related standards

• NFPA 70E
• IEEE 1584-2018

Read the document

Publication details

Authors

Nicolaus Jennings; David Wetz; Alexander Johnston; Rick Langley; Nancy LaFlair; John Heinzel

Publisher

IEEE Open Journal of Industry Applications (Institute of Electrical and Electronics Engineers (IEEE))

Year

2025

Volume

6

Pages

689-702

Document language

English

Source type

Scientific

Access

Open licence — downloadable

DOI

10.1109/OJIA.2025.3616287

Licence

CC BY 4.0 — share with attribution

Licence evidence

ieeexplore.ieee.org

Citation