In the context of accelerating global energy transition, the safe operation of large-scale energy storage stations faces severe challenges. Thermal runaway (TR)-induced arc faults, characterized by ultrahigh temperatures and energy density, have emerged as a critical factor exacerbating fire and explosion risks in battery energy storage system. When the electrical safety clearance between components is less than the critical breakdown distance, insulation failure caused by high temperature or mechanical damage can lead to gas dielectric discharge. The high-temperature gases, particles, and electrolyte ejected during thermal runaway can significantly reduce insulation strength and alter the local dielectric environment. Structural degradation caused by loose electrical connections or chemical corrosion can lead to insulation damage and evolve into a continuous arc. We have observed that the arc temperature can reach up to 6000 K, far exceeding the melting point of steel, aluminum, and copper commonly used in battery energy storage system. Therefore, an in-depth understanding of the arc generation mechanism in battery energy storage system is crucial for improving system safety.
1. Arc Formation Mechanisms
The formation of arcs in battery energy storage system is a typical multi-factor coupling process. The core mechanisms can be summarized into three physical pathways. First, insulation failure induced type: battery deformation, mechanical vibration, or assembly stress leads to structural displacement, exposing bare metal. When the safe distance between exposed conductors and adjacent components decreases to the critical breakdown distance, gas breakdown occurs, forming a conductive channel. Second, thermal runaway product modification type: high-temperature gases and particles ejected during battery thermal runaway change the local dielectric environment. Combustible electrolyte vapors such as ethylene carbonate (EC) and diethyl carbonate (DEC) reduce insulation strength to 20% to 50% of that of air. Metal particles form a gas-solid mixed medium, inducing non-uniform field distortion. Third, structural degradation accumulation type: during long-term operation, loose electrical connections or chemical corrosion cause insulation damage, eventually producing partial discharge in high-voltage circuits, gradually evolving into continuous arcs.
The following table summarizes the three main types of arc formation mechanisms in battery energy storage system.
| Arc Type | Trigger Condition | Physical Mechanism | Consequence |
|---|---|---|---|
| Insulation Failure Induced | Thermal/mechanical stress, battery swelling | Electric field exceeds dielectric strength | Gas breakdown, short circuit |
| TR Product Modified | Ejected gas, particles, electrolyte | Reduced insulation strength, field distortion | Flashover, arc ignition |
| Structural Degradation | Corrosion, loose connections | Gradual insulation breakdown | Partial discharge, continuous arcing |
2. Insulation Material Failure Leading to Arc
In battery energy storage system, insulation material failure is one of the core causes of arc faults. As battery energy density continues to increase and voltage platforms upgrade, insulation materials face more severe electrical-thermal-mechanical stress coupling challenges. We conducted high-temperature tests on common insulation materials used in battery energy storage system, including mica paper, thermal insulation bubble film, structural adhesive, and battery blue film. The insulation materials were placed in a box furnace, heated to 400 °C at a rate of 10 °C/min for 1 minute, then allowed to cool naturally. Mica paper showed little change in shape at 400 °C. The thermal insulation bubble film and battery blue film showed minor shape changes at 300 °C but obvious shrinkage deformation at 400 °C. The structural adhesive began to pyrolyze at 200 °C, softened and lost its original shape at 300 °C, and completely melted at 400 °C.
The resistance values of these materials at different temperatures are summarized in the table below.
| Temperature (°C) | Mica Paper (Ω) | Thermal Bubble Film (Ω) | Structural Adhesive (Ω) | Blue Film (Ω) |
|---|---|---|---|---|
| 200 | 3.04 × 109 | 2.94 × 109 | 1.40 × 1012 | 3.17 × 109 |
| 300 | 2.91 × 1011 | 1.21 × 1011 | — | 1.73 × 1010 |
| 400 | 5.20 × 1011 | 1.59 × 1011 | — | 1.40 × 1011 |
After high-temperature exposure, the resistance values of the insulation materials changed. However, the shrinkage or damage caused by high temperatures can remove insulation protection from live parts, increasing the risk of short circuits and arcing.
3. Particle-Induced Arc
During thermal runaway, batteries eject large amounts of high-temperature combustible gases and solid particles. When these particles flow with gas and deposit between high-voltage components in the system, they can induce arcs. We have investigated the ejected particles from batteries with different cathode materials, capacities, and triggering methods. The studies focused on particle mass, size distribution, elemental composition, chemical type, and morphology. For example, we have shown that particles from Li(Ni0.8Co0.1Mn0.1)O2 lithium-ion batteries can induce arcs with a critical breakdown voltage as low as (99 ± 5) V at an electrode gap of 1 mm. At a 4 mm electrode gap, the breakdown voltage is (155 ± 5) V, which is only 1.2% of the air breakdown voltage.
In one study, we designed a device to simulate particle-induced arcs from battery thermal runaway. The device included a DC power supply (voltage U0), an arc generation area with electrodes at a certain distance (area voltage U, resistance R), a load resistor R0, an ammeter, and a high-speed data acquisition system. Particles were filled in the electrode gap to simulate the actual conditions after battery ejection. We investigated the relationship between particle size, electrode gap, circuit resistance, and arc critical voltage.
The critical breakdown voltage for particle-induced arcs can be expressed by the following empirical relationship.
$$ U_c \ge 4.08L^2 $$
where Uc is the critical voltage and L is the electrode gap distance. This equation provides a safety boundary for arc prevention design in battery systems.
In a study of particle-filled gaps, particles with approximately 70% graphite content significantly reduced the air insulation strength, lowering the breakdown voltage to 0.9% to 2.3% of that of pure air. For electrode gaps between 1 mm and 8 mm, the arc breakdown voltage showed a positive quadratic correlation with the gap distance. When the gap exceeded 8 mm, particles could not induce arcs below 400 V. The breakdown voltage showed a negative correlation with particle size. Larger particles (size > 100 μm) were more likely to induce arcs. The load resistance had little effect on the breakdown voltage.
The following table summarizes the experimental conditions and key findings for particle-induced arcs.
| Parameter | Value/Range | Observation |
|---|---|---|
| Electrode gap (mm) | 1 – 8 | Arcing possible, breakdown voltage increases with gap |
| Particle size (μm) | >100 | More likely to induce arcs |
| Metal fragment size (mm) | >2 | Significantly lowers breakdown voltage |
| Breakdown voltage (V) | 99 – 155 (for 1 – 4 mm gap) | Only 1.2% of air breakdown voltage |
4. Electrolyte-Induced Arc
In battery energy storage system, LiFePO4 batteries are commonly used. Their thermal runaway ejecta contain a large amount of electrolyte and mixtures of electrolyte and particles. Therefore, the mechanism of electrolyte-induced arc is critical for the safety of battery energy storage system. We designed an experiment to simulate the situation where the battery casing is eroded by ejected electrolyte while subjected to high voltage. Using aluminum electrodes, we studied the arc behavior with electrolyte as the medium.
The electrical conductivity of the electrolyte was first measured.
| Sample | Measurement 1 (mS/cm) | Measurement 2 (mS/cm) | Measurement 3 (mS/cm) | Average (mS/cm) |
|---|---|---|---|---|
| Electrolyte | 13.29 | 13.54 | 13.52 | 13.45 |
The conductivity of the electrolyte is approximately 13.45 mS/cm. Compared to copper (5.8 × 108 mS/cm) and aluminum (3.5 × 108 mS/cm) at room temperature, the electrolyte cannot directly conduct under high voltage to form a short circuit path. When a voltage of 235 V was applied, a brief flash of light appeared between the electrodes and electrolyte for about 1 second, but no stable arc occurred initially. As the voltage continued, the electrolyte rapidly evaporated due to the high temperature from the high voltage, producing white smoke. After 18 to 28 seconds, the electrolyte transformed into a black paste-like substance. Then, between 28 and 30 seconds, arc-like flashes appeared on the surface of the black paste. At 30 seconds, a sudden, intense arc occurred, accompanied by a dazzling light. The temperature at the measurement point rapidly increased to approximately 890 °C between 30 and 39 seconds. The temperature of the arc itself was estimated to be around 6000 °C. After the arc ceased, a weak flame persisted for about 1 second. The aluminum electrodes showed severe ablation and partial melting.
The process of electrolyte-induced arcing involves the following steps:
- Evaporation of electrolyte under high voltage.
- Formation of a conductive path through carbonized residue.
- Sudden breakdown and arc formation.
- High-temperature ablation of electrodes.
5. Harm of Induced Arc
In battery energy storage system, multiple battery cells and modules are connected in series and parallel to meet power output requirements. The stability of these connections is crucial for system safety. During cycling, factors such as battery aging, swelling, or mechanical stress can cause solder joints connecting batteries to loosen or break. This can lead to arc formation at the connection points, which can cause severe electrochemical corrosion, shorten battery life, and trigger thermal runaway chain reactions leading to fire.
We constructed an experimental platform to simulate arc faults in battery systems. The main components included a DC power supply, arc triggering device, battery, and electronic load. The arc generation device had one end fixed to the battery via an insulating clamp and the other end with an electrode tip fixed to a movable slider. A controller regulated the stepper motor speed to control the electrode movement and the distance between the electrode and battery.
We investigated the effect of state of charge (SOC) on arc behavior. The arc evolution in batteries at different SOC levels can be divided into four stages: pre-arc initiation, arc evolution, arc extinction, and battery thermal runaway. In the early arc stage, the arc appeared as bright blue-white light. The arc ignited the insulation material on the battery cap, producing flames. For batteries at 0% and 30% SOC, after arc extinction, white smoke began to leak from the negative terminal area. This is due to irreversible reactions from the sustained high temperature of the arc. The insulation sealing material near the negative terminal was severely burned, creating holes that allowed internal gas to leak. For batteries at 60% and 100% SOC, flame ejection occurred near the negative terminal before arc extinction. The arc ignited the combustible gas, intensifying the damage to the LiFePO4 battery. The higher the SOC, the shorter the time required for the arc to induce battery failure, and the more severe the consequences such as flame spread, mass loss, and swelling.
The effect of circuit voltage was also studied. At a loop current of 20 A and a separation gap of 0.5 mm, the arc duration was 2.8 seconds for 40 V, 4.4 seconds for 42 V, and 7.1 seconds for 50 V. The higher the voltage, the longer the arc duration.
The effect of circuit current was investigated. At a voltage of 200 V and a gap of 0.5 mm, the arc duration was 20.3 seconds for 20 A, 5.7 seconds for 30 A, and 4.4 seconds for 40 A. Higher current led to shorter arc duration because the higher power consumed more material to sustain the arc.
The effect of gap distance was also examined. At a voltage of 200 V and a current of 20 A, the arc duration was 20.6 seconds for a 0.5 mm gap and 13.2 seconds for a 1.0 mm gap. A larger gap required more energy to maintain the arc, resulting in a shorter duration.
The following table summarizes the effects of various parameters on arc duration.
| Parameter | Change | Effect on Arc Duration |
|---|---|---|
| State of Charge (SOC) | Increase | Decrease |
| Circuit Voltage (Udc) | Increase | Increase |
| Circuit Current (Idc) | Increase | Decrease |
| Gap Distance (L) | Increase | Decrease |
These results indicate that SOC, voltage, current, and electrode gap all significantly influence the severity and duration of arcs. The findings provide a basis for designing safer battery energy storage system.
6. Arc Modeling and Simulation
Currently, a fully coupled model for arc breakdown in high-voltage systems after battery thermal runaway is still lacking. Arc simulation typically uses nonlinear differential equations or other mathematical methods. By providing input values and boundary conditions through assumptions and simplifications, numerical methods are used to solve arc models, obtaining parameters such as electric field, magnetic field, temperature field, and fluid field. The core of arc simulation is to accurately describe the electro-thermal-fluid multi-physics coupling behavior of the plasma.
Existing models can be classified into three categories based on modeling principles and applicable scales. First, black-box/empirical models, such as the Cassie and Mayr models, are simple to calculate but can only describe external arc characteristics, ignoring internal complex processes. Second, magnetohydrodynamic (MHD) models, which consider the interaction between electromagnetic and fluid dynamics equations. Third, micro-particle models, which provide a more detailed description of particle behavior.
For MHD models, the governing equations are based on the conservation of mass, momentum, and energy, coupled with Maxwell’s equations. The basic set of MHD equations for arc simulation can be expressed as:
$$ \frac{\partial \rho}{\partial t} + \nabla \cdot (\rho \mathbf{u}) = 0 $$
$$ \rho \frac{\partial \mathbf{u}}{\partial t} + \rho (\mathbf{u} \cdot \nabla)\mathbf{u} = -\nabla P + \nabla \cdot \tau + \mathbf{J} \times \mathbf{B} $$
$$ \rho c_p \frac{\partial T}{\partial t} + \rho c_p \mathbf{u} \cdot \nabla T = \nabla \cdot (k \nabla T) + \sigma E^2 – Q_{rad} $$
where ρ is density, u is velocity, P is pressure, τ is the viscous stress tensor, J is current density, B is magnetic flux density, cp is specific heat capacity, T is temperature, k is thermal conductivity, σ is electrical conductivity, E is electric field, and Qrad is radiative heat loss.
In battery system arc simulation, researchers have built finite element models for series arc faults based on MHD equations. These models analyze arc voltage at different electrode gaps and study the distribution of electric field, magnetic field, arc temperature, and flow velocity around the arc. The models determine the maximum values and locations of these parameters, analyzing the electro-thermal coupling characteristics of the arc and the multi-physics field distribution.
The following table compares the different arc modeling approaches.
| Model Type | Principles | Advantages | Limitations |
|---|---|---|---|
| Black-box (Cassie/Mayr) | Empirical differential equations | Simple, computationally efficient | Ignores internal physical processes |
| Magnetohydrodynamic (MHD) | Coupled fluid dynamics and electromagnetics | Captures multi-physics coupling | Complex, high computational cost |
| Micro-particle | Detailed particle and plasma physics | High accuracy for specific phenomena | Very complex, limited applicability |
Current research on thermal runaway arc models in battery energy storage system is still in its early stages. The specific assumptions of traditional models lead to deviations from actual operating conditions. Given the large number of electrical connection points in battery energy storage system and the complex factors such as high temperature, ejecta, and dynamic changes in safety clearance during thermal runaway, building a multi-physics coupled “thermal-electrical-mechanical-chemical” model for arc hazards faces significant challenges.
7. Conclusion and Perspective
In this paper, we have systematically analyzed the mechanism of arc formation in battery energy storage system, reviewed the research progress on thermal runaway-induced arc hazards from the perspectives of insulation failure, ejection medium, arc-battery interaction, and modeling. The key findings can be summarized as follows. First, high-temperature environments cause insulation materials such as structural adhesive to soften and fail at temperatures above 300 °C. This can reduce the electrical safety clearance below the critical breakdown distance, directly triggering an arc. Second, ejected particles from batteries can reduce the breakdown voltage to 0.9% to 2.3% of that of pure air. Studies have shown that particles larger than 100 μm and metal fragments are more likely to induce arcs. Electrolyte leakage can trigger a chain reaction of deflagration under high voltage, increasing the risk of thermal runaway propagation. Third, the numerous electrical connectors inside the battery system are prone to arc faults due to battery aging and swelling. These arcs can even induce battery thermal runaway, increasing the probability of catastrophic events.
Current studies on arc formation and simulation in battery energy storage system are often based on specific conditions and assumptions, which differ from actual arc events. Future research should focus on reproducing real accident scenarios, both experimentally and through modeling, to better understand the mechanism of arc formation in battery energy storage system and clarify the development pattern of the battery thermal runaway-arc hazard chain. This will strengthen the safety protection of battery energy storage system and reduce the probability of catastrophic accidents.