The global transition towards renewable energy sources like wind and solar has created an urgent demand for large-scale energy storage solutions to stabilize power grids. Electrochemical energy storage, particularly lithium-ion battery technology, has become the dominant choice due to its high energy density, long cycle life, and declining costs. Energy storage power stations, composed of thousands to millions of individual battery cells, are now critical infrastructure. However, the safety of these systems, specifically the risk of thermal runaway (TR), remains a paramount concern. Thermal runaway is a chain of exothermic reactions within a cell that can lead to rapid temperature rise, gas venting, fire, and even explosion. In the densely packed configuration of an energy storage station, a single cell’s thermal runaway can propagate to neighboring cells, potentially escalating into a catastrophic failure of the entire system. Therefore, developing reliable early detection methods and effective safety prevention and control technologies is fundamental to the safe and widespread deployment of lithium-ion battery energy storage.
This article systematically reviews the current state of research in this critical area. It begins by analyzing the fundamental processes of thermal runaway in energy storage cells, including its triggers and characteristic parameter evolution. It then delves into the latest state detection technologies, covering traditional and advanced sensor-based methods, electrochemical impedance spectroscopy (EIS), and ultrasonic testing. Subsequently, the safety prevention and control strategies implemented at the energy storage power station level are summarized. Finally, the article concludes with a discussion on existing challenges and future research directions.
1. Fundamental Process of Thermal Runaway in Energy Storage Cells
Lithium iron phosphate (LFP) batteries are the predominant chemistry used in stationary energy storage in many regions due to their superior thermal stability compared to high-nickel cathodes. Understanding the thermal runaway process in these energy storage cells is the first step towards effective prevention.
1.1 Triggers of Thermal Runaway
Thermal runaway can be initiated by three primary abuse conditions: thermal abuse (e.g., external heating), electrical abuse (e.g., overcharge, short circuit), and mechanical abuse (e.g., crush, nail penetration). In the context of energy storage power stations, thermal abuse from faulty thermal management and electrical abuse from system control failures are the most common triggers. The core mechanism involves a vicious cycle where heat generation from side reactions exceeds the cell’s heat dissipation capacity, leading to a rapid, uncontrolled temperature increase.
- Thermal Abuse: When the cell temperature exceeds a critical threshold (e.g., ~90°C for LFP), internal exothermic side reactions begin to dominate. Factors like higher State of Charge (SOC) and aging lower the onset temperature and accelerate the process.
- Electrical Abuse – Overcharge: Overcharging forces excess lithium ions to plate as metallic lithium (dendrites) on the anode. These dendrites can pierce the separator, causing an internal short circuit, which generates intense localized Joule heating and triggers thermal runaway. Higher charging rates (C-rates) exacerbate this process.
1.2 Internal Reaction Process and Heat Propagation
The thermal runaway process involves a sequence of irreversible exothermic reactions:
- SEI Decomposition (~90-120°C): The Solid Electrolyte Interphase (SEI) layer on the anode decomposes.
$$ \text{(CH}_2\text{OCO}_2\text{Li)}_2 \rightarrow \text{Li}_2\text{CO}_3 + \text{C}_2\text{H}_4 + \text{CO}_2 + 0.5\text{O}_2 $$ - Anode Reaction with Electrolyte (>120°C): The exposed lithiated anode (e.g., LixC6) reacts with the electrolyte solvents (e.g., EC, DEC).
$$ 2\text{Li} + \text{C}_3\text{H}_4\text{O}_3 \text{(EC)} \rightarrow \text{Li}_2\text{CO}_3 + \text{C}_2\text{H}_4 $$
$$ 2\text{Li} + \text{C}_5\text{H}_{10}\text{O}_3 \text{(DEC)} \rightarrow \text{Li}_2\text{CO}_3 + \text{C}_4\text{H}_{10} $$ - Cathode Decomposition and Electrolyte Reaction (~180°C for LFP): The cathode material releases oxygen, which reacts with the electrolyte.
$$ (1-x)\text{LiFePO}_4 + x\text{FePO}_4 \rightarrow (1-x)\text{LiFePO}_4 + \frac{x}{4}\text{Fe}_2\text{P}_2\text{O}_7 + \frac{x}{4}\text{O}_2 $$ - Electrolyte Decomposition and Binder Reaction (>200°C): Further decomposition of electrolyte and PVDF binder occurs, releasing various flammable and toxic gases.
Once triggered in one cell, heat propagates to adjacent cells via conduction, radiation, and convection of ejected hot materials, potentially leading to cascading thermal runaway across a module or rack.
1.3 Evolution of Characteristic Parameters
The internal reactions manifest as distinct changes in external parameters, which form the basis for detection.
1.3.1 Temperature
Temperature is the most direct indicator. The thermal runaway process can be divided into stages using characteristic temperatures (θ1, θ2, θ3). θ1 is the self-heating onset temperature. θ2 is the thermal runaway trigger point, often linked to separator meltdown. θ3 is the peak temperature during the violent reaction phase. For large-format LFP energy storage cells, θ1 is typically below 100°C, θ2 ranges from 100-200°C, and θ3 can exceed 500°C.
| Capacity (Ah) | θ₁ (°C) | θ₂ (°C) | θ₃ (°C) |
|---|---|---|---|
| 25 | 70 | 120 | 263 |
| 60 | 86.6 | 131 | 514 |
| 280 | 70.6 | 200.7 | 340.7 |
1.3.2 Voltage
Voltage behavior depends on the trigger. During overcharge, voltage rises until a sudden drop to near zero at internal short circuit. Under thermal abuse, voltage typically shows a gradual decline as the separator deteriorates, followed by an abrupt drop to zero. Abnormal voltage plateaus or rapid drops can serve as early warning signs of internal faults like micro-shorts.
1.3.3 Gas Generation
The decomposition reactions produce characteristic gases. Detection of these gases offers a very early warning opportunity, often before significant temperature rise. The primary gases from LFP energy storage cells include H2, CO, CO2, and various hydrocarbons (C2H4, CH4, etc.). Hydrogen (H2) is particularly significant as it is often generated early from reactions between plated lithium and the binder.
| Key Gases Detected | Typical Stage | Significance for Detection |
|---|---|---|
| CO, CO₂, H₂ (Early) | SEI decomposition & initial anode reaction | Earliest warning signal, often pre-vent. |
| C₂H₄, CH₄ (Mid-stage) | Severe electrolyte decomposition | Indication of advancing thermal abuse. |
| HF, POF₃ (Late) | Electrolyte salt (LiPF₆) decomposition | Toxic gas release, confirms severe failure. |
2. Thermal Runaway State Detection Technologies
Moving from understanding the process to actively monitoring it, a variety of detection technologies are employed or under research for energy storage power stations.
2.1 Conventional Monitoring in Prefabricated Compartments
Current energy storage stations primarily rely on a Battery Management System (BMS) monitoring voltage, current, and temperature, combined with compartment-level smoke and temperature sensors. While crucial for operational control, these systems often detect thermal runaway only at a late stage (e.g., after venting or fire ignition), limiting early warning capability.
2.2 Sensor-Based Detection Technologies
Advanced sensor strategies aim to detect anomalies earlier and more reliably.
| Detection Category | Method | Principle | Advantages | Challenges |
|---|---|---|---|---|
| Ex-situ (External) | Voltage/Temperature Threshold | Alarm on predefined limits (e.g., dT/dt > X °C/min). | Simple, easy to implement. | High false alarms, late detection. |
| Model-Based | Compares real measurements with model predictions; large residuals indicate fault. | Good real-time performance, incorporates physics. | Requires accurate model; sensitive to noise and aging. | |
| Data-Driven (AI) | Uses ML algorithms (LSTM, CNN, clustering) to identify fault patterns from historical data. | No explicit model needed; robust to complex patterns. | Requires vast, high-quality training data; “black-box” nature. | |
| Ex-situ (External) | Gas Sensing | Detects specific gases (e.g., H₂, CO) released early during TR. | Potential for very early warning. | Sensor cross-sensitivity, placement optimization, calibration. |
| In-situ (Internal) | Embedded Sensors | Micro-sensors (FBG, NTC, thin-film) implanted inside the cell to measure core temperature/strain. | Direct, fast, and accurate internal state measurement. | Invasive; impacts cell integrity & cost; long-term reliability. |
2.3 Electrochemical Impedance Spectroscopy (EIS)
EIS is a non-invasive technique that probes the internal electrochemical state of an energy storage cell by applying a small AC current/voltage over a range of frequencies and measuring the impedance response. The impedance spectrum, often modeled by an equivalent circuit, reflects parameters like ohmic resistance (RΩ), charge transfer resistance (Rct), and diffusion elements (Ww).
The impedance at a given frequency is calculated as:
$$ Z(f) = \frac{U_{\text{amp}}(f)}{I_{\text{amp}}(f)} e^{j\theta(f)} $$
where $U_{\text{amp}}$ and $I_{\text{amp}}$ are the AC voltage and current amplitudes, and $\theta$ is the phase shift.
Changes in internal state (temperature, SOC, SOH, failure onset) alter these equivalent circuit parameters. For example, the phase shift ($\theta$) in the mid-frequency range (10-100 Hz) is highly sensitive to internal temperature and relatively independent of SOC, enabling internal temperature estimation. Recent research focuses on using single-frequency or multi-frequency dynamic impedance monitoring to detect the onset of internal short circuits or overcharge conditions in real-time, providing a promising pathway for early fault diagnosis in energy storage cells.
2.4 Ultrasonic Testing Technology
This emerging technique uses ultrasonic waves to probe the internal mechanical state of an energy storage cell. As ultrasonic waves propagate through the cell’s layered structure, their time-of-flight (TOF) and amplitude are influenced by the material properties of the electrodes and electrolyte.
- State Detection: Changes in SOC and SOH affect electrode modulus and density, altering ultrasonic signal features. This can be used for state estimation.
- Safety Monitoring: The formation of gas bubbles (from electrolyte decomposition or plating) and internal deformations (swelling) cause significant scattering and attenuation of ultrasonic waves. This sensitivity allows for the very early detection of failure mechanisms like lithium plating or initial gas generation, potentially offering warning well before thermal runaway.
While highly promising for its sensitivity to internal physical changes, interpreting ultrasonic signals requires disentangling the effects of temperature, SOC, and specific failure modes, often requiring integration with other sensor data.
3. Safety Prevention and Control Technologies for Energy Storage Power Stations
When detection indicates a failure, active measures are required to contain it. Safety strategies operate at different scales within a station.
3.1 Centralized Suppression (Compartment-Level)
This is the traditional approach where a fire detection system triggers the release of a suppression agent to flood the entire container or compartment. The goal is to extinguish flames and cool the cells to break the thermal runaway chain reaction.
| Suppressant Type | Example | Advantages | Disadvantages |
|---|---|---|---|
| Water-Based | Water Mist | Excellent cooling, non-toxic, low cost. | Conductivity risk, may not prevent re-ignition. |
| Gas-Based | Perfluorohexanone (FK-5-1-12) | Clean, non-conductive, effective flame suppression. | Moderate cooling, higher cost, environmental concerns. |
| Gas-Based | HFC-227ea (FM200) | Clean, fast-acting. | Poor cooling, produces HF at high temps. |
Research shows that water mist offers the best cooling, while synthetic clean agents like perfluorohexanone provide fast flame knockdown. A combined strategy, e.g., using gas for rapid flame suppression followed by water mist for sustained cooling, is often considered optimal.
3.2 Thermal Runaway Propagation Inhibition
This strategy focuses on preventing a single cell’s thermal runaway from spreading to its neighbors, thereby containing the event. This is achieved through thermal management and physical barriers.
- Thermal Interface Materials: Using materials with optimized thermal conductivity between cells can help dissipate heat during normal operation but slow down heat transfer during a runaway event.
- Phase Change Materials (PCMs): PCMs integrated into the battery pack absorb large amounts of heat during the melting process, effectively delaying the temperature rise of adjacent cells.
- Physical Barriers & Spacing: Incorporating fire-resistant insulation materials or increasing the gap between cells/modules reduces radiant and conductive heat transfer.
3.3 Distributed and Targeted Suppression
The most advanced strategy involves targeted intervention at the smallest possible unit.
- Rack/Cluster-Level: Suppression systems dedicated to a single battery rack or cluster, allowing localized agent release without affecting the entire station.
- Module-Level: Nozzles or agents dedicated to a single module.
- Cell-Level: The ultimate targeted approach. Examples include:
- Intrusive Nozzles: Directed into the module towards cell venting paths.
- Self-Triggering Devices: Passive devices placed near cell vents that activate (e.g., through thermal decomposition) when exposed to hot vent gases, releasing suppressant directly onto the failing cell.
Distributed suppression minimizes agent use, reduces collateral damage, and can act faster, but increases system complexity and cost.
| Control Strategy | Scale | Key Advantage | Key Challenge |
|---|---|---|---|
| Centralized Suppression | Entire Compartment | Simple architecture, easier maintenance. | Slow response, massive agent use, collateral damage. |
| Propagation Inhibition | Inter-Cell/Module | Prevents escalation, passive safety. | Adds weight/volume, limited to delaying propagation. |
| Distributed/Targeted Suppression | Rack, Module, or Cell | Fast, precise, minimizes impact. | High complexity, cost, and need for reliable detection. |
4. Conclusion and Future Perspectives
Ensuring the safety of lithium-ion battery energy storage power stations is a multi-layered challenge requiring deep understanding, early detection, and effective suppression. Research has made significant progress in mapping the thermal runaway process of large-format energy storage cells, developing advanced detection algorithms (model-based and AI-driven), and exploring novel monitoring techniques like EIS and ultrasonics. Safety engineering has evolved from compartment-flooding towards smarter, more targeted propagation inhibition and suppression systems.
However, critical challenges remain. Future research should focus on the following key directions:
- Fundamental Understanding: Establishing more precise correlations between internal chemical/ physical changes during early failure and the corresponding external multi-parameter signatures (voltage, impedance, ultrasound, gas) is crucial for developing reliable early warning algorithms.
- Advanced In-situ Sensing: Developing robust, miniaturized, and cost-effective sensors that can be integrated into energy storage cells to provide direct internal state data (temperature, pressure, gas) without compromising cell performance or longevity.
- Standardization and Benchmarking: The lack of standardized testing protocols and performance benchmarks for detection and suppression systems makes comparison and validation difficult. International efforts are needed in this area.
- Integration and System-Level Design: Future energy storage systems should be designed with safety as a core principle, integrating detection sensors, thermal management, and suppression hardware at the module/pack level from the outset. This includes developing intelligent Battery Management Systems (BMS) that can fuse data from multiple detection sources (voltage, temperature, gas, impedance) to make accurate and timely safety decisions.
- Next-Generation Cell Chemistries and Designs: Ultimately, improving the inherent safety of the energy storage cell itself is paramount. Research into more stable electrolytes (e.g., solid-state), less reactive electrode materials, and cell designs with built-in safety features (e.g., current interrupt devices, advanced separators) will provide the foundation for safer energy storage stations.
By addressing these challenges through continued interdisciplinary research and development, the safety and reliability of lithium-ion battery energy storage power stations can be significantly enhanced, supporting their critical role in the global transition to a sustainable energy future.