The energy storage industry has reached a new peak of development. Electrochemical energy storage systems, prized for their excellent frequency regulation and peak-shaving capabilities, are now trending toward large-scale application. Lithium iron phosphate (LiFePO4) energy storage cells are widely used in energy storage power stations due to their long cycle life and high energy density. However, as these cells evolve toward higher capacity and power, thermal runaway incidents continue to occur, with several major safety accidents reported globally, posing a severe threat to public safety. A critical safety concern in such stations is the vertically arranged battery modules. Following the thermal runaway of a lower-layer energy storage cell, the ejected flammable gases can ignite, and the resulting flame can induce fire propagation to the upper-layer cells. Understanding this vertical fire propagation is therefore paramount for system safety.
While extensive research exists on the thermal runaway characteristics of individual LiFePO4 cells and the horizontal thermal propagation within modules, studies on vertical fire propagation between layers remain insufficient. In particular, there is a lack of quantitative analysis of the energy accumulation process in the upper-layer energy storage cell before fire propagation is triggered, and the energy transfer mechanisms during this triggering process are not fully clarified. To address this knowledge gap, my research focuses on investigating the fire propagation characteristics and energy transfer mechanisms within double-layer energy storage battery modules. I designed experiments using 100 Ah LiFePO4 energy storage cells to simulate a critical scenario and unravel the underlying physics.
The core of my experimental design involved three sets of fire propagation tests using double-layer modules. The number of energy storage cells in each layer was varied systematically: one cell per layer, two cells per layer, and three cells per layer. In all tests, the bottom cell in the first row was heated to induce venting. The ejected gases were actively ignited to simulate a real-fire scenario following cell failure. I meticulously recorded the experimental phenomena and the temperature evolution of each energy storage cell using strategically placed thermocouples. By analyzing the temperature rise rates and identifying distinct heating stages, I aimed to characterize the fire propagation behavior. Crucially, by comparing results from the different module configurations, I was able to quantitatively deconstruct the cumulative energy transferred from the bottom cells to a top energy storage cell during the triggering process and decouple the heat contribution from different paths (bottom vs. side surfaces).
Experimental Methodology and Design
The research object was a commercially available prismatic LiFePO4 energy storage cell designed for stationary storage. The key parameters of this energy storage cell are summarized in Table 1.
| Parameter | Value |
|---|---|
| Nominal Capacity | 100 Ah |
| Mass | (2245 ± 5) g |
| Dimensions (L×W×H) | 130 mm × 36 mm × 211.8 mm |
| Positive Electrode | LiFePO₄ |
| Negative Electrode | Graphite |
| Nominal Voltage | 3.2 V |
| Upper Cut-off Voltage | 3.65 V |
| Lower Cut-off Voltage | 2.5 V |
| State of Charge (SOC) for test | 100% |
| In-plane Thermal Conductivity | 21.24 W/(m·°C) |
Table 1: Parameters of the energy storage cell under investigation.
All cells were conditioned to 100% SOC using a standard protocol: constant-current discharge at 1C to 2.5 V, followed by a 1-hour rest, and then constant-current-constant-voltage charge at 0.5C to 3.65 V until the current dropped to 0.05C. The experiments were conducted in a standard combustion chamber equipped with an exhaust fan for safety. The double-layer module was placed on a stainless-steel frame. For the thermal triggering, an 800 W heating film was attached to the large face of the first bottom energy storage cell. Upon gas venting from this cell, a pulse igniter was used to ignite the gases immediately. The thermal runaway of an energy storage cell was defined as the moment when its temperature on the surface opposite the heater exhibited a temperature rise rate exceeding 5 °C/s for three consecutive data points concurrent with a voltage drop below 1 V, at which point the heater was manually turned off.
The experimental matrix, designed to probe the threshold for vertical fire propagation, is shown in Table 2. Thermocouples were attached to the front (heated face), side, back, and vent port of each energy storage cell. For the top-layer cells, an additional thermocouple was positioned on their bottom surface to measure the impinging flame temperature. The modules were assembled with 5 cm of vertical spacing between the layers, representing a severe but plausible scenario. Mica plates and aluminum alloy fixtures were used to minimize heat loss from the module ends.
| Test Group | Configuration | Fire Propagation Outcome |
|---|---|---|
| Test 1 (Non-propagation) | 1 cell per layer | No |
| Test 2 (Non-propagation) | 2 cells per layer | No |
| Test 3 (Fire-propagation) | 3 cells per layer | Yes |
Table 2: Experimental design for investigating fire propagation thresholds in double-layer energy storage modules.
Fire Propagation Phenomena and Temperature Characteristics
The experimental observations revealed distinct phases leading to fire propagation. In Tests 1 and 2 (non-propagation groups), the thermal runaway of the bottom-layer energy storage cells produced sustained flames, but the top-layer cells did not undergo thermal runaway. In Test 1, the single top energy storage cell was heated but only reached a peak average temperature of 120.9 °C. In Test 2, the heating from two bottom cells was more intense, causing one of the top energy storage cells to vent, but not enough to trigger a full thermal runaway. The key findings emerged from Test 3, where the sequential thermal runaway of three bottom energy storage cells ultimately triggered the simultaneous thermal runaway of all three top-layer energy storage cells.
The temperature profiles for the fire-propagation test (Test 3) are critical for understanding the mechanism. The bottom energy storage cells (#1, #2, #3) underwent thermal runaway sequentially, with characteristic jet fires during their most violent stage. The top energy storage cells (#4, #5, #6) exhibited a markedly different and more severe thermal response. Their peak average temperature reached 640.2 °C, which was 115.9 °C (22.1%) higher than the peak temperature of the bottom cells. Furthermore, the maximum temperature rise rate for the top energy storage cells was 14.0 °C/s, significantly exceeding the 7.5 °C/s peak rate observed in the bottom cells—an increase of 86.7%. This indicates that an energy storage cell experiencing fire propagation from below is subjected to more intense heating, combining its internal exothermic reactions with external flame impingement, leading to a more hazardous event.
A detailed analysis of the temperature curve for a top energy storage cell (e.g., cell #5) prior to its own thermal runaway reveals a clear step-wise heating process, as visualized in the concept of temperature rise stages. This heating profile directly correlates with the events occurring in the bottom layer. Each bottom energy storage cell contributed two distinct heating phases to the cells above: a Flame Baking Stage initiated when the bottom cell vented and its gases burned steadily, and a more intense Flame Jet Stage initiated when that bottom cell underwent violent thermal runaway with a high-velocity jet fire. Consequently, the top energy storage cell experienced three sets of these stages. The quantitative comparison is striking:
| Heating Stage | Average Duration | Average Temp. Rise | Average Temp. Rise Rate |
|---|---|---|---|
| Flame Baking Stage | ~237 s | ~27.0 °C | ~6.9 °C/min |
| Flame Jet Stage | ~117 s | ~30.3 °C | ~15.6 °C/min |
Table 3: Comparison of heating stages for a top energy storage cell during fire propagation triggering.
While both stages caused a similar magnitude of temperature increase (27-30 °C), the rate of heating during the Flame Jet Stage was approximately 2.3 times faster than during the Flame Baking Stage. This highlights the dramatically higher heating efficiency of the jet fire produced by a bottom energy storage cell in full thermal runaway, which is a key driver in the fire propagation process.
Quantitative Energy Transfer Mechanism
The step-wise temperature rise of the top energy storage cell points directly to a cumulative energy transfer process. To quantify this, I utilized the data from the non-propagation tests (Tests 1 and 2). In these tests, the thermal events in the bottom layer were isolated in time, allowing for a clean measurement of the energy imparted to the top layer by a specific number of failing bottom energy storage cells. The energy accumulated within a top energy storage cell, \( Q_{acc} \), can be calculated from its average temperature rise, \(\Delta T_{ave}\), its mass \( m \), and its specific heat capacity \( c \):
$$ Q_{acc} = c \cdot m \cdot \Delta T_{ave} $$
Using \( c = 985.3 \, \text{J/(kg·°C)} \) and \( m = 2.24 \, \text{kg} \), the accumulated energy in a top energy storage cell was calculated for each test scenario. This energy accumulation was then compared to characteristic energy thresholds derived from Accelerating Rate Calorimetry (ARC) data for a single energy storage cell, namely the energy required to reach the venting temperature \( Q_{vent} \) and the energy required to trigger thermal runaway \( Q_{TR} \).
$$ Q_{vent} = c \cdot m \cdot (T_{vent} – T_{initial}) $$
$$ Q_{TR} = c \cdot m \cdot (T_{TR} – T_{initial}) $$
The results of this energy accounting are summarized in Table 4.
| Scenario (Bottom Cells Active) | Energy in Top Cell, \( Q_{acc} \)** | Relation to ARC Thresholds | Observed Outcome |
|---|---|---|---|
| 1 Bottom Cell | 249.1 kJ | \( Q_{acc} < Q_{vent} \) (84.6% of \( Q_{vent} \)) | Heating only, no venting. |
| 2 Bottom Cells | 334.3 kJ | \( Q_{vent} < Q_{acc} < Q_{TR} \) (97.8% of \( Q_{TR} \)) | Venting triggered, but no thermal runaway. |
| 3 Bottom Cells (Triggering) | 379.7 kJ | \( Q_{acc} \approx Q_{TR} \) (111% of \( Q_{TR} \)) | Thermal runaway (Fire Propagation). |
Table 4: Cumulative energy accumulation in a top energy storage cell and its correlation with fire propagation triggering.
The data reveals a clear triggering boundary. The fire propagation energy, \( Q_{trig} \), required to cause thermal runaway in a top energy storage cell under these conditions was approximately 379.7 kJ. This value aligns closely with the independently measured \( Q_{TR} \) from ARC tests (341.8 kJ), with the difference attributable to heat losses not present in the ideal ARC environment. This confirms that the fundamental mechanism for fire propagation is the sequential addition of energy from failing bottom energy storage cells until the critical threshold of the top energy storage cell is surpassed.
Decoupling Heat Transfer Pathways
Understanding which surfaces of the top energy storage cell receive the most heat is vital for designing effective countermeasures. The total triggering energy \( Q_{trig} \) is supplied through two main paths: heat flux through the bottom surface \( Q_{bo} \) impinged by the flame, and heat flux through the side surfaces \( Q_{side} \) exposed to the general fire plume. The energy balance at the point of triggering can be simplified as:
$$ Q_{trig} = Q_{bo} + Q_{side} $$
where self-heating and environmental heat losses prior to thermal runaway are negligible. The heat transferred through the bottom surface can be estimated using the recorded temperature difference between the cell’s bottom surface \( T_{bo}(t) \) and its core/average temperature \( T_{ave}(t) \), the cell’s in-plane thermal conductivity \( \lambda \), the bottom surface area \( A_{bo} \), and the conduction distance \( \delta \):
$$ Q_{bo} = \frac{\lambda \cdot A_{bo}}{\delta} \int_{t_0}^{t_{trig}} \left( T_{bo}(t) – T_{ave}(t) \right) dt $$
Using the data from Test 3 and performing this calculation allows us to decouple the contributions. The results provide crucial insight for safety engineering:
| Heat Transfer Pathway | Transferred Energy | Percentage of \( Q_{trig} \)** |
|---|---|---|
| Through Bottom Surface (\( Q_{bo} \)) | 180.5 kJ | 47.5% |
| Through Side Surfaces (\( Q_{side} \)) | 199.2 kJ | 52.5% |
| Total (\( Q_{trig} \)) | 379.7 kJ | 100% |
Table 5: Decoupled energy contribution from different heat transfer paths to a top energy storage cell.
The finding that heat transfer through the sides contributed slightly more (52.5%) than through the bottom (47.5%) is significant. It indicates that the entire exposed surface of the energy storage cell is vulnerable during a fire propagation event. Relying solely on bottom-side insulation would be insufficient; a holistic approach to isolating the cell from external fire impingement is necessary.
Synthesized Energy Transfer Mechanism and Implications
My research synthesizes a clear mechanism for vertical fire propagation in double-layer energy storage modules. The process is governed by the cumulative and step-wise energy transfer from sequentially failing bottom energy storage cells to the cells directly above. Each bottom cell failure adds a discrete packet of energy through combined flame baking and jet fire impingement. The top energy storage cell integrates this energy until its internal energy content surpasses the critical thermal runaway threshold, at which point it fails catastrophically, often propagating the fire further. The entire sequence for triggering a top-layer energy storage cell is as follows: the first bottom cell brings it to ~65% of the needed energy; the second pushes it past the venting point to ~88%; and the third finally supplies the remaining ~12% to cross the thermal runaway boundary.
This understanding has direct and critical implications for the safety design of energy storage systems:
- Horizontal Propagation Control is Primary: Since vertical fire propagation is contingent on the sequential thermal runaway of multiple bottom-layer energy storage cells, preventing or delaying horizontal thermal runaway within a layer is the first and most effective defense. Implementing robust inter-cell thermal barriers within a module is essential.
- The “Golden Time” for Intervention: The fire propagation process is not instantaneous. My experiments show a significant time window (over 1000 seconds in the tested configuration) between the initial bottom cell failure and the eventual thermal runaway of the top energy storage cells. This provides a crucial opportunity for fire detection and suppression systems to activate and intervene before a full-scale module-to-module fire develops.
- Comprehensive Cell Isolation: The nearly equal contribution of side and bottom heating to the triggering energy means that energy storage cells must be protected on all exposed faces. Module and enclosure designs should incorporate insulating materials or spacing not just below, but also around the sides of cells to mitigate heat flux from adjacent failing cells or external fires.
In conclusion, my investigation into the fire propagation characteristics of double-layer energy storage modules reveals a quantifiable, energy-driven mechanism. The severity of top-layer failure is greater than that of the initiating cells. The triggering process is defined by distinct, cumulative heating stages correlated with bottom-cell failure events, with jet fires being particularly effective heaters. I have quantitatively defined the fire propagation energy threshold for the studied energy storage cell and decoupled the significant heat transfer contributions from both the bottom and side surfaces. These findings provide a scientific foundation and clear engineering guidance for designing safer, more resilient energy storage systems through targeted thermal management, timely active protection, and comprehensive cell isolation strategies.