Energy storage power stations play a pivotal role in the utilization of renewable energy sources such as solar and wind power. Their safe operation is critical for achieving global energy structure transformation and meeting carbon neutrality targets. Among various electrochemical storage technologies, lithium-ion batteries dominate due to their high energy density, long cycle life, and efficiency. However, the widespread deployment of these systems has been accompanied by a series of safety incidents, including fires, explosions, and toxic gas releases, which pose significant threats to personnel, property, and the environment. In this review, we systematically analyze the key safety risks associated with lithium-ion battery energy storage power stations and summarize recent advances in early warning technologies and fire suppression methods. Special emphasis is placed on the performance and protection of the energy storage cell, which is the fundamental unit of any battery system. By understanding the failure mechanisms and implementing multi-layered safety strategies, we can enhance the reliability and safety of energy storage cell systems. This article provides insights into risk mitigation from technical, managerial, and operational perspectives, aiming to support the sustainable growth of the energy storage cell industry.
1. Safety Risks of Energy Storage Power Stations
Energy storage power stations integrate thousands of energy storage cell units, power conversion systems, and thermal management components. The primary risks can be classified into four categories: fire hazards, gas generation, electric shock, and waste battery recycling issues.
1.1 Fire Hazards
Lithium-ion batteries use organic solvents with low boiling points in their electrolytes. Under abnormal conditions such as overcharging, over-discharging, internal short circuits, or thermal runaway, these solvents can ignite and lead to catastrophic fires. Statistical data from the Electric Power Research Institute shows that more than 80 energy storage accidents occurred globally in the past decade, most involving lithium-ion batteries. The following table summarizes several representative incidents from 2017 to 2022, illustrating the diverse causes of fire events.
| Date | Location | Capacity | Battery Type | Incident Description | Cause |
|---|---|---|---|---|---|
| 2022.02 | Moss Landing, USA | 300 MW/1.2 GWh | NMC | About 10 battery racks melted | Overcharging induced thermal runaway |
| 2022.01 | Uiseong, South Korea | 1.5 MWh | NMC | Solar plant ESS fire | Unknown |
| 2021.04 | Beijing, China | 25 MWh | LFP | Explosion during operation | Overcharging induced thermal runaway |
| 2020.07 | Victoria, Australia | 450 MWh | NMC | Tesla Megapack fire | Coolant leakage during testing |
| 2019.04 | Arizona, USA | 2 MW/2.47 MWh | NMC | Fire killed 4 firefighters | Internal cell defect |
| 2017.12 | Shanxi, China | 9 MW/4.5 MWh | LFP | Second fire in same station | Internal short circuit in one cell |
The thermal runaway of an energy storage cell can be described by the heat balance equation, where the heat generation rate exceeds the heat dissipation rate, leading to a self-accelerating temperature rise. A simplified lumped model is:
$$ \rho C_p \frac{dT}{dt} = \dot{Q}_{gen} – \dot{Q}_{diss} $$
where \(\rho\) is density, \(C_p\) is specific heat capacity, \(T\) is temperature, \(\dot{Q}_{gen}\) is the heat generation from exothermic reactions, and \(\dot{Q}_{diss}\) is the heat dissipated to the environment. The heat generation term during abuse can be expressed as an Arrhenius-type equation:
$$ \dot{Q}_{gen} = \sum_i A_i \exp\left(-\frac{E_{a,i}}{RT}\right) H_i $$
where \(A_i\) is the pre-exponential factor, \(E_{a,i}\) is the activation energy, \(R\) is the gas constant, and \(H_i\) is the enthalpy of the \(i\)-th reaction (e.g., SEI decomposition, electrolyte decomposition). These equations underline the importance of early detection to prevent runaway.
1.2 Gas Generation
During abnormal operation, lithium-ion energy storage cells release flammable and toxic gases due to electrolyte decomposition and side reactions. Typical gases include hydrogen (H₂), carbon monoxide (CO), methane (CH₄), and ethylene (C₂H₄). The composition varies with battery chemistry. The following table lists the gas species and their approximate volume percentages for different cathode materials during thermal runaway:
| Gas | LFP (%) | NMC (%) | LCO (%) |
|---|---|---|---|
| CO₂ | 40–50 | 35–45 | 30–40 |
| H₂ | 20–30 | 20–25 | 15–20 |
| CO | 10–20 | 15–25 | 20–30 |
| CH₄ | 5–10 | 5–10 | 5–8 |
| C₂H₄ | 2–5 | 3–6 | 3–5 |
Hydrogen and carbon monoxide are considered the most reliable early warning indicators because they appear before visible smoke or flame. The accumulation of flammable gases can lead to explosions if ignited. The lower explosion limit (LEL) for H₂ is 4% in air, and for CO it is 12.5%. Continuous monitoring of these gases is therefore essential for energy storage cell safety.
1.3 Electric Shock
Large-scale energy storage stations operate at high voltage levels, typically from 10 kV to 220 kV, depending on capacity. The series-parallel connection of thousands of energy storage cells creates high-voltage DC buses. Improper isolation during maintenance, inadequate grounding, or operator error can result in severe electric shock accidents. The risk increases with the system voltage, which can be expressed as:
$$ V_{total} = N_s \times V_{cell} $$
where \(N_s\) is the number of cells in series and \(V_{cell}\) is the nominal voltage of a single energy storage cell (e.g., 3.2 V for LFP, 3.7 V for NMC). For a 1.5 kV DC bus, \(N_s \approx 470\) cells for LFP. Strict adherence to safety protocols and the use of insulation monitoring devices are mandatory.
1.4 Waste Battery Recycling
After 3–5 years of service, retired lithium-ion energy storage cells contain valuable metals (Li, Co, Ni, Mn, Cu) as well as toxic substances (electrolyte solvents, fluorine compounds). Improper disposal leads to soil and water contamination and poses health risks. The following table summarizes the potential hazards of key components in a typical energy storage cell:
| Component | Hazard |
|---|---|
| Lithium | Heart and thyroid dysfunction |
| Cobalt | Allergic dermatitis, carcinogenic potential |
| Nickel | Carcinogenic, toxic to aquatic life |
| Fluorine | Soil/water contamination, respiratory damage |
| Phosphorus | Neurotoxicity, carcinogenic |
| Copper | Digestive system damage |
Recycling not only mitigates environmental harm but also recovers critical materials. The economic incentive is high: the metal content in spent energy storage cells is 10–100 times higher than in natural ores. Proper end-of-life management is thus an integral part of safety for the whole lifecycle.
2. Research Progress on Safety Protection Measures
To address the safety challenges, extensive research has focused on early warning technologies and fire suppression methods. These measures aim to detect anomalies at the earliest possible stage and to mitigate the consequences if a failure occurs.
2.1 Early Warning Technologies
Early warning systems monitor various parameters of the energy storage cell and its environment to predict impending failures. The main approaches include electrical parameter monitoring, gas sensing, smoke/sound detection, and intelligent algorithms.
2.1.1 Electrical Parameter-Based Warning
The battery management system (BMS) continuously measures voltage, current, and internal resistance of each energy storage cell. Abnormal changes—such as a sudden voltage drop, resistance increase, or current imbalance—can indicate internal short circuits or overcharge conditions. The state-of-health (SOH) and state-of-charge (SOC) are estimated using models like the equivalent circuit model (ECM). A simple ECM is given by:
$$ V_t = V_{oc} – IR_0 – V_{RC} $$
where \(V_t\) is terminal voltage, \(V_{oc}\) is open-circuit voltage, \(I\) is current, \(R_0\) is internal resistance, and \(V_{RC}\) is the voltage across the RC network representing polarization. Advanced BMS now integrate cloud computing to process vast datasets from thousands of energy storage cells, enabling predictive maintenance.
2.1.2 Gas Sensing Warning
As shown in Table 2, H₂ and CO are produced early in the degradation process before thermal runaway. Metal-oxide semiconductor (MOS) sensors can detect H₂ at concentrations as low as 10 ppm. Experimental studies demonstrate that H₂ is captured 639 seconds earlier than smoke and 769 seconds earlier than fire in LFP/graphite cells. A dual-gas detection system using H₂ and CO improves reliability. The sensor output voltage \(V_s\) can be related to gas concentration \(C\) by:
$$ V_s = V_0 \left(1 + k C^n\right) $$
where \(V_0\) is baseline voltage, \(k\) and \(n\) are constants. Calibration is essential to avoid false alarms.
2.1.3 Smoke and Sound Sensing
When the safety valve of an energy storage cell opens due to internal pressure, it produces an audible hissing sound and releases aerosolized electrolyte. Smoke detectors and acoustic sensors can capture these signals, but they are slower than gas sensors. To reduce false alarms, advanced algorithms like YOLOv3-based image recognition are being developed to identify vapor-liquid ejecta. However, these techniques are still in laboratory stages.
2.1.4 Intelligent Warning Systems
Machine learning and artificial intelligence enable real-time analysis of multi-physical parameters. For instance, an energy storage monitoring system (ESMS) collects temperature, voltage, current, and humidity from each energy storage cell module and transmits data wirelessly to a central processor. By training neural networks on historical failure data, the system can predict anomalies hours before they occur. The architecture includes data acquisition, feature extraction, and classification layers. A typical loss function for anomaly detection is:
$$ \mathcal{L} = \frac{1}{N} \sum_{i=1}^N \left( y_i – \hat{y}_i \right)^2 + \lambda \|w\|^2 $$
where \(y_i\) is the true label (0=normal, 1=abnormal), \(\hat{y}_i\) is the predicted probability, and \(\lambda\) is the regularization parameter. This approach is transforming safety management for large-scale energy storage cell installations.
2.2 Fire Suppression Methods
Once a fire occurs, rapid suppression is critical to prevent propagation. Suppression strategies include material improvements, thermal management, and active firefighting.
2.2.1 Cell Material Optimization
Enhancing the intrinsic safety of the energy storage cell is the most fundamental approach. Solid-state electrolytes, for example, eliminate flammable liquid solvents. However, commercial viability is still limited. For current liquid-electrolyte cells, flame-retardant additives such as tris(1,3-dichloroisopropyl) phosphate (TDCPP) and methyl phosphonic acid dimethyl ester (DMMP) are used. Adding 5–10% TDCPP can prevent electrolyte ignition. Additionally, cathode coating strategies—e.g., boron-gradient doping in NMC—improve thermal stability. The activation energy for oxygen release can be increased, as described by:
$$ k = A \exp\left(-\frac{E_a}{RT}\right) $$
where a higher \(E_a\) reduces reaction rates at elevated temperatures.
Lithium iron phosphate (LFP) energy storage cells are inherently safer than NMC due to their olivine structure. The image above illustrates a typical LFP battery module used in energy storage applications. Such cells have lower energy density but superior thermal stability, making them preferred for large-scale stations where safety is paramount.
2.2.2 Thermal Management Optimization
Efficient thermal management keeps the operating temperature of each energy storage cell within 25–50°C and minimizes temperature differences between cells (≤5°C). Active air cooling is common, but immersion liquid cooling is gaining traction for its superior heat transfer. The heat transfer coefficient for immersion cooling can be expressed by:
$$ \dot{Q}_{cool} = h A (T_{cell} – T_{fluid}) $$
where \(h\) is the convective heat transfer coefficient, \(A\) is the cell surface area, and \(T_{fluid}\) is the coolant temperature. With proper flow design, the inter-cell temperature difference can be reduced to under 3°C. Intelligent thermal management systems using predictive control further improve safety.
2.2.3 Firefighting Techniques
Two main types of fire suppression systems are used: gas-based and water-based. Common gas agents include heptafluoropropane (HFC-227ea) and perfluorohexanone (Novec 1230). The required extinguishing concentration for heptafluoropropane is around 8% by volume in air. For Novec 1230, it is about 4–6%. These agents are clean and leave no residue. In severe cases, large-scale water deluge systems can prevent thermal runaway propagation by providing massive cooling. The following table compares the key properties:
| Agent | Concentration (vol%) | Extinguishing time (s) | Environmental impact (GWP) |
|---|---|---|---|
| Heptafluoropropane | 8–10 | <10 | 3220 |
| Perfluorohexanone | 4–6 | <5 | 1 |
| Water (deluge) | Large volume | Continuous | None |
The selection depends on station size, ventilation, and environmental regulations. Combining early warning with automatic suppression creates a multi-level defense for energy storage cell systems.
3. Countermeasures and Recommendations
To comprehensively improve the safety level of lithium-ion battery energy storage power stations, we propose the following strategies, which address design, standards, monitoring, and emergency response.
- Optimize energy storage station design. During the planning phase, comply with regulations for site selection, capacity sizing, battery type selection, and fire protection layout. The thermal runaway propagation model of energy storage cell modules should be simulated to ensure adequate spacing and fire barriers.
- Strengthen safety standards and management. China has over 20 national standards for electrochemical energy storage, but specific standards for fire risk (like IEC 62485-5 or NFPA 855) are lacking. Establishing dedicated safety codes for energy storage cell systems is urgent. Regular safety training for operators is also essential.
- Develop intelligent early warning systems. Invest in multi-parameter sensing (temperature, voltage, gas, acoustic) and AI-based analytics to predict failures at the energy storage cell level. A centralized monitoring center should integrate data from all cells and provide real-time alerts.
- Establish emergency response mechanisms. Formulate contingency plans jointly with power, fire, environmental, and medical departments. Conduct regular drills. For energy storage cell fires, consider both clean agent and water suppression strategies. A tiered response based on severity ensures minimal damage.
The following table summarizes the recommended safety measures across different stages of the energy storage cell lifecycle:
| Stage | Measure | Key Technology/Activity |
|---|---|---|
| Design | Cell selection | Use LFP or solid-state cells for higher safety |
| Design | Module layout | Thermal barriers, spacing to prevent propagation |
| Operation | BMS monitoring | Real-time voltage, temperature, gas detection |
| Operation | Thermal management | Immersion cooling or advanced air cooling |
| Operation | Fire suppression | Novec 1230 automatic system + water deluge |
| End-of-life | Recycling | Hydrometallurgical or pyrometallurgical recovery |
4. Conclusion
Lithium-ion battery energy storage power stations are indispensable for integrating renewable energy into the grid, but their safety risks—especially fire and explosion—remain a major barrier to large-scale deployment. In this article, we have systematically reviewed the safety risks associated with energy storage cell systems, including fire, gas generation, electric shock, and waste management. We then presented state-of-the-art early warning technologies (electrical, gas, acoustic, and intelligent) and fire suppression methods (material improvement, thermal management, and firefighting agents). Finally, we proposed a set of countermeasures covering design, standards, monitoring, and emergency response. By implementing these multi-layered strategies, the safety of energy storage cell installations can be significantly enhanced, fostering the sustainable development of the global energy storage market.