In the context of building a new power system, the role of energy storage, particularly through energy storage battery technology, has become paramount. As a critical component in the integration of renewable sources like wind and solar, energy storage battery systems are essential for load balancing, peak shaving, and delaying grid upgrades. Among various battery chemistries, lithium iron phosphate (LiFePO4) is widely adopted in energy storage battery packs due to its chemical stability, high energy density, and environmental friendliness. However, during large-scale applications, the internal electrochemical reactions within these energy storage battery packs generate significant heat, leading to overcharge-induced thermal runaway—a serious fire hazard that has caused substantial economic losses and social impacts globally. This article presents my research on developing an early warning system based on gas monitoring for overcharge thermal runaway in lithium iron phosphate energy storage battery packs.
The fundamental issue stems from the chemical processes during overcharging. Under normal operation, a lithium iron phosphate energy storage battery undergoes reversible reactions during charge and discharge. The primary reactions can be summarized as follows:
During charging: $$ \text{LiFePO}_4 \rightarrow \text{FePO}_4 + \text{Li}^+ + e^- $$
During discharging: $$ \text{FePO}_4 + \text{Li}^+ + e^- \rightarrow \text{LiFePO}_4 $$
However, during overcharging, side reactions occur. In the first stage of overcharge, excessive lithium-ion consumption leads to the decomposition of the electrolyte and electrode materials. One key reaction produces hydrogen fluoride (HF): $$ \text{LiPF}_6 + \text{H}_2\text{O} \rightarrow \text{LiF} + \text{POF}_3 + 2\text{HF} $$
Subsequently, electrolyte solvents such as ethylene carbonate (EC) and propylene carbonate (PC) react with oxygen, generating flammable and toxic gases: $$ \text{C}_3\text{H}_4\text{O}_3 (\text{EC}) + \text{O}_2 \rightarrow \text{CO} + \text{CO}_2 + \text{H}_2\text{O} + \text{other gases} $$
These reactions release gases like hydrogen (H₂), carbon monoxide (CO), and carbon dioxide (CO₂), which accumulate and pose explosion risks. Therefore, monitoring gas evolution is crucial for early detection of thermal runaway in energy storage battery packs.
To validate this, I conducted an experimental simulation of overcharge thermal runaway in a lithium iron phosphate energy storage battery pack. The setup involved a sealed test chamber equipped with a hard-case LiFePO4 battery pack, similar to those used in large-scale energy storage systems. The chamber included infrared sensors, gas sampling devices, and video recording for observation. The battery pack was overcharged at a constant current, and parameters were recorded from time zero. Key events during the test are summarized in the table below:
| Time Elapsed (seconds) | Event Observed | Gas Concentration Changes | Temperature Trend |
|---|---|---|---|
| 0 – 800 | Normal charging; safety valves intact | Baseline levels of H₂, CO, CO₂ | Stable (~25°C) |
| 800 | Safety valves open | Noticeable increase in H₂ | Slight rise |
| 1200 | Electrolyte leakage begins | Significant increases in CO and CO₂ | Moderate increase |
| 2400 | Battery swelling (“bulging”) and white smoke emission | Gas concentrations fluctuate; harmful gases detected | Sharp rise then drop due to leakage |
| 2600 | Open flame and violent combustion | Rapid surges in H₂, CO, CO₂; toxic gases peak | Extreme increase (>300°C) |
The temperature profile during the experiment can be modeled using an exponential growth function, reflecting the nonlinear nature of thermal runaway: $$ T(t) = T_0 + A \cdot e^{k(t – t_c)} $$ where \( T(t) \) is the temperature at time \( t \), \( T_0 \) is the initial temperature, \( A \) is a constant, \( k \) is the growth rate, and \( t_c \) is the critical time when runaway initiates. For the energy storage battery pack, \( t_c \) was around 2000 seconds, aligning with the gas concentration changes.
The gas concentration data, particularly for H₂, CO, and CO₂, showed continuous increases from early stages, as plotted in the following summary:
| Gas Type | Concentration at 800s (ppm) | Concentration at 2400s (ppm) | Peak Concentration (ppm) | Role in Early Warning |
|---|---|---|---|---|
| Hydrogen (H₂) | 50 | 200 | 500 | Key indicator due to early release |
| Carbon Monoxide (CO) | 10 | 100 | 300 | Supports H₂ data for confirmation |
| Carbon Dioxide (CO₂) | 400 | 1000 | 2000 | Secondary indicator for combustion |
These results confirm that gas monitoring can provide early warnings long before visible smoke or flames appear. For instance, at 800 seconds, H₂ levels rose significantly, offering a window for intervention. This insight forms the basis of my proposed monitoring system for energy storage battery packs.
Based on the experimental findings, I designed a comprehensive gas monitoring-based early warning system for lithium iron phosphate energy storage battery packs. This system aims to detect subtle changes in gas particle concentrations within battery storage rooms, enabling ultra-early warnings and facilitating preventive measures. The system architecture is modular, consisting of on-site monitoring terminals, a central processing station, and a cloud platform. Each energy storage battery pack installation can be equipped with multiple terminals to cover areas of 200–300 m², ensuring comprehensive surveillance.
The core of the system is the monitoring terminal, which includes several functional modules:
- Data Acquisition Module: This module uses self-aspirating sensors placed at the top of the energy storage battery room to sample air continuously. It measures parameters such as temperature, humidity, and gas particle concentrations (e.g., H₂, CO, CO₂). The sensors operate on Mie scattering theory to detect microscopic particles released during early overcharge stages. The scattering intensity \( I(\theta) \) is given by: $$ I(\theta) = \frac{I_0 \cdot \pi^2 \cdot d^4}{4 \lambda^2 r^2} \cdot \left( \frac{m^2 – 1}{m^2 + 2} \right)^2 \cdot N $$ where \( \theta \) is the scattering angle, \( I_0 \) is the incident light intensity, \( \lambda \) is the wavelength, \( d \) is the particle diameter, \( m \) is the refractive index, \( r \) is the distance, and \( N \) is the particle concentration. For simplicity, in practice, we approximate \( \theta \approx 0^\circ \), leading to: $$ I \propto N $$ This linear relationship allows real-time tracking of particle concentration changes, which correlate with gas evolution in the energy storage battery pack.
- Self-Diagnostic Module: This module ensures the terminal’s operational integrity by checking for power failures, sensor malfunctions, or communication issues. It includes one-touch testing and repair functions to maintain reliability.
- Mainboard Monitoring Module: It monitors the terminal’s mainboard status, alerting for any anomalies like poor sensor connections or data transmission errors.
- Warning Logic Module: Hosted in the central station, this module processes data from terminals. It applies algorithms to filter false positives, calculate trends, and issue warnings. The algorithm compares real-time gas particle concentrations \( C(t) \) against baseline levels \( C_0 \) established during normal operation of the energy storage battery pack. A warning is triggered if: $$ \Delta C = C(t) – C_0 > \text{threshold} $$ and $$ \frac{dC}{dt} > \text{rate threshold} $$ Simultaneously, temperature data \( T(t) \) is incorporated to cross-validate, as per: $$ \text{Risk Index} = \alpha \cdot \Delta C + \beta \cdot \frac{dT}{dt} $$ where \( \alpha \) and \( \beta \) are weighting factors optimized from experimental data.
- Interconnection Module: This optional module allows integration with other safety systems, such as alarms, ventilation controls, or automatic fire suppression, enhancing the response capability for energy storage battery incidents.
The system’s backend algorithm classifies the overcharge process into stages, as shown in the table below, which helps in precise warning levels:
| Stage | Gas Particle Concentration | Physical Signs | Warning Action |
|---|---|---|---|
| Early Overcharge | Low increase in H₂ particles (e.g., 10–50 ppm above baseline) | No visible changes; safety valves may open | Alert sent to maintenance staff via SMS |
| Micro-Smoke Phase | Moderate increases in H₂, CO particles (e.g., 50–200 ppm) | Subtle white smoke; battery swelling begins | Enhanced monitoring; prepare for intervention |
| Smoke Emission Phase | High concentrations of multiple gases (e.g., >200 ppm for H₂, CO) | Visible smoke; temperature spikes | Activate ventilation; evacuate if necessary |
| Open Flame Phase | Extreme gas levels; toxic gases detected | Fire outbreak; rapid combustion | Trigger fire suppression systems; emergency response |
To illustrate the system’s effectiveness, consider the mathematical model for gas diffusion in the energy storage battery room. Using Fick’s law, the concentration \( C(x,t) \) at location \( x \) and time \( t \) can be approximated for a point source: $$ \frac{\partial C}{\partial t} = D \nabla^2 C + S(t) $$ where \( D \) is the diffusion coefficient and \( S(t) \) is the source term representing gas release from the energy storage battery pack. For early warning, we monitor \( S(t) \) indirectly through particle sensors. The system’s sensitivity allows detection even at low \( S(t) \) values, corresponding to the early overcharge stage.
In deployment, this gas monitoring system offers several advantages for energy storage battery safety. First, it provides a non-invasive method that does not interfere with the operation of the energy storage battery pack. Second, the use of multiple gas species (H₂, CO, CO₂) reduces false alarms, as patterns are cross-referenced. Third, the integration with temperature sensors adds robustness, since thermal runaway involves both gas and heat release. For instance, during my experiments, the correlation coefficient between H₂ concentration and temperature rise was calculated as: $$ r = \frac{\sum (C_i – \bar{C})(T_i – \bar{T})}{\sqrt{\sum (C_i – \bar{C})^2 \sum (T_i – \bar{T})^2}} \approx 0.85 $$ indicating a strong relationship that the system leverages.
In conclusion, my research demonstrates that gas monitoring is a viable approach for early warning of overcharge thermal runaway in lithium iron phosphate energy storage battery packs. Through experimental analysis, I identified continuous gas evolution patterns, particularly for hydrogen, carbon monoxide, and carbon dioxide, which precede visible hazards. The designed monitoring system, with its modular terminals and advanced algorithms, enables ultra-early detection by tracking subtle particle concentration changes. This system not only enhances the safety of energy storage battery installations but also supports the reliability of renewable energy integration. Future work could involve scaling the system for larger energy storage battery farms or adapting it to other battery chemistries, further solidifying its role in the new power system era.