As the global energy transition accelerates at an unprecedented pace, the domain of electrical energy storage has emerged as a cornerstone of the modern energy infrastructure. Among various electrochemical storage technologies, lithium-ion batteries have become a dominant force, powering applications ranging from grid-scale storage systems to residential energy solutions. In our research, we focus on the critical intersection of performance longevity and post-aging safety characteristics of large-format lithium iron phosphate (LFP) batteries, which are widely deployed in stationary energy storage applications. The ability to predict battery life under realistic operating conditions and to assess safety risks after prolonged cycling is of paramount importance for the reliable and secure operation of energy storage systems. Our work systematically investigates the accelerated aging behavior under elevated temperature conditions and evaluates the thermal safety evolution of batteries after extensive cycling.
Lithium-ion batteries, while offering high energy density and long cycle life, undergo complex degradation mechanisms that are accelerated by thermal stress. The degradation pathways include solid electrolyte interphase (SEI) film growth, loss of active lithium, structural degradation of electrode materials, and electrolyte decomposition. These mechanisms not only reduce the available capacity but also fundamentally alter the thermal stability of the battery, potentially increasing the risk of thermal runaway events. In the context of energy storage battery applications, where systems are expected to operate reliably for 10 to 15 years, understanding the coupling between aging and safety is essential. Our study addresses this gap by establishing quantitative relationships between cycling temperature, capacity fade, and post-cycling thermal runaway behavior.
The experiments were conducted using commercial large-format prismatic LFP cells sourced from multiple manufacturers. This multi-source approach ensures that our findings are representative of the broader technology landscape rather than being specific to a single production process. The nominal capacity of the cells used for cycling tests ranged from 280 Ah to 345 Ah, with nominal energy between 800 Wh and 1000 Wh. For safety characterization, cells with a nominal capacity of 314 Ah and nominal energy between 600 Wh and 700 Wh were employed. All cells shared the LFP/graphite chemistry, which is the predominant chemistry for stationary energy storage due to its intrinsic safety advantages and long cycle life.
The cycling performance tests were conducted following the protocols outlined in the GB/T 36276 standard for lithium-ion batteries used in electrical energy storage. Two temperature conditions were selected: 25 °C, representing typical room temperature operation, and 45 °C, representing elevated temperature conditions that accelerate aging processes. For each temperature condition and for each of the five battery models, two cells were tested, resulting in a total of ten cells for the cycling study. The cycling protocol consisted of constant power charging at the rated power Prc to the upper cutoff voltage, followed by a 10-minute rest period, and then constant power discharging at the rated power Prd to the lower cutoff voltage. This charge-discharge cycle was repeated for 1000 cycles. Capacity retention was monitored periodically to quantify the aging progression.
The safety performance evaluation employed thermal runaway testing following the GB/T 36276 standard. For this purpose, ten different battery models from five manufacturers were selected, with two cells per model (twenty cells total) divided equally between pre-cycling and post-cycling tests. The post-cycling cells underwent 1000 cycles at 45 °C before thermal runaway testing. The thermal runaway test involved placing the cell in a controlled environment, attaching heating elements and temperature sensors to the cell surface, and applying a constant current charge while heating the cell. The onset of thermal runaway was defined as the point when three consecutive temperature rise rates exceeded 3 °C/s, or when fire or explosion occurred. The temperature at which thermal runaway was triggered was recorded as the critical thermal runaway temperature.
Cycling Performance at Different Temperatures
The cycling performance results at 25 °C are presented in the following summary. For all five battery models, an initial increase in capacity was observed during the first few cycles, a phenomenon attributed to the gradual activation of electrode materials and improved electrolyte wetting during the early stages of cycling. This activation effect allowed the internal electrochemical reactions to proceed more efficiently, temporarily boosting the measured capacity above the nominal value. Following this initial rise, a gradual capacity decline ensued as degradation mechanisms began to dominate. After 1000 cycles at 25 °C, the capacity retention values for the five models were 95.82%, 95.81%, 96.24%, 95.63%, and 95.06%, yielding an average retention of 95.71%. These results demonstrate the excellent cycling stability of LFP chemistry under moderate temperature conditions.
In contrast, the cycling performance at 45 °C exhibited markedly different characteristics. The initial capacity rise observed at 25 °C was largely suppressed at the elevated temperature. This suppression occurs because the accelerated degradation processes at higher temperatures, including enhanced SEI growth and increased side reactions, counteract the activation effects from the very beginning of cycling. The capacity decline was substantially faster at 45 °C compared to 25 °C. After 1000 cycles, the capacity retention values for the five models were 90.18%, 90.26%, 90.24%, 91.88%, and 91.47%, with an average of 90.81%. The accelerated capacity fade at elevated temperature is primarily driven by the temperature dependence of the parasitic reactions that consume active lithium and degrade electrode materials.
To quantify the acceleration effect of temperature on aging, we analyzed the ratio of capacity degradation rates between the two temperature conditions. The capacity degradation rate at each cycle interval is defined as:
$$D_{cap}(n) = 1 – \frac{Q(n)}{Q_0}$$
where Q(n) is the capacity after n cycles and Q_0 is the initial capacity. The ratio of the degradation rates at 45 °C to that at 25 °C was calculated and analyzed as a function of cycle number. During the initial cycles, this ratio was relatively high due to the initial capacity increase observed at 25 °C, which temporarily reduced the apparent degradation rate at room temperature. As cycling progressed and the initial activation effect subsided, the ratio decreased and asymptotically approached a value of 2. This convergence indicates that after the initial stabilization period, the aging rate at 45 °C is approximately twice that at 25 °C.
The relationship between the degradation rate ratio and cycle number was modeled using an inverse function:
$$R(n) = a + \frac{b}{n}$$
where R(n) is the ratio of degradation rates, n is the cycle number, and a and b are fitting parameters. For our data, the fitting yielded a = 0.4 and b = 22.7, with a correlation coefficient R² = 0.9976, indicating excellent agreement between the model and experimental observations. The asymptotic value of R(n) as n approaches infinity is a = 2.0, confirming that the long-term aging acceleration factor at 45 °C relative to 25 °C is 2. This finding provides a quantitative basis for predicting the lifetime of energy storage batteries under different temperature regimes. Specifically, if the cycling profile and capacity fade characteristics at a reference temperature are known, the lifetime at elevated temperatures can be estimated using this acceleration factor, enabling more accurate lifetime projections for systems operating in warm climates or under thermal management constraints.
The implications of this acceleration factor for the energy storage battery industry are substantial. System designers can use this relationship to optimize thermal management strategies, balancing the cost of cooling against the benefits of extended battery life. For instance, maintaining battery temperature below 35 °C rather than allowing operation at 45 °C could potentially double the calendar life, although the exact benefit depends on the specific temperature profile and the activation energy of the dominant degradation pathways.
We also examined the variability among the five battery models. Despite differences in cell design, electrode formulation, and manufacturing processes, the capacity retention values at each temperature condition were remarkably consistent. At 25 °C, the standard deviation of capacity retention after 1000 cycles was less than 0.5%, while at 45 °C, the standard deviation was approximately 0.8%. This consistency suggests that the LFP chemistry, combined with mature manufacturing processes, yields highly reproducible aging behavior across different suppliers. For end-users, this uniformity translates to predictable system performance and facilitates the development of generalized lifetime models that can be applied across multiple battery sources.
The detailed cycling data for each battery model at both temperatures are summarized in the table below.
| Battery Model | Temperature (°C) | Initial Capacity (Ah) | Capacity After 1000 Cycles (Ah) | Capacity Retention (%) |
|---|---|---|---|---|
| Model A | 25 | 298.5 | 286.0 | 95.82 |
| Model B | 25 | 312.3 | 299.2 | 95.81 |
| Model C | 25 | 335.8 | 323.2 | 96.24 |
| Model D | 25 | 280.4 | 268.2 | 95.63 |
| Model E | 25 | 305.6 | 290.5 | 95.06 |
| Average | 25 | — | — | 95.71 |
| Model A | 45 | 298.5 | 269.2 | 90.18 |
| Model B | 45 | 312.3 | 281.9 | 90.26 |
| Model C | 45 | 335.8 | 303.0 | 90.24 |
| Model D | 45 | 280.4 | 257.6 | 91.88 |
| Model E | 45 | 305.6 | 279.5 | 91.47 |
| Average | 45 | — | — | 90.81 |
The capacity degradation rate analysis is further refined by examining the differential capacity fade as a function of cycle number. We define the incremental degradation rate as:
$$\Delta D_{cap}(n) = D_{cap}(n) – D_{cap}(n-1)$$
This metric reveals how the aging rate evolves over time. At 25 °C, the incremental degradation rate decreases during the first 100 cycles as the initial activation phase transitions into steady-state aging. Beyond cycle 200, the incremental degradation rate stabilizes at approximately 0.004% per cycle, indicating a linear aging regime. At 45 °C, the incremental degradation rate is initially higher and remains elevated throughout the cycling test, stabilizing at approximately 0.008% per cycle after cycle 200, which is consistent with the acceleration factor of 2 observed in the cumulative degradation ratio.
The linear aging regime observed after the initial cycles is characteristic of LFP batteries operating under moderate conditions, where the dominant degradation mechanism is the gradual consumption of active lithium due to continuous SEI growth and repair. The SEI layer, while protecting the anode from direct electrolyte reduction, undergoes constant restructuring and thickening, consuming lithium ions in the process. The rate of SEI growth is temperature-dependent, following an Arrhenius-type relationship. The activation energy for SEI growth in LFP systems has been reported in the literature to be in the range of 40 to 60 kJ/mol, which would yield an acceleration factor of approximately 2 to 3 for a temperature increase from 25 °C to 45 °C, consistent with our observed factor of 2.
Thermal Runaway Behavior of Aged Energy Storage Batteries
The safety implications of battery aging were investigated through thermal runaway testing of cells before and after 1000 cycles at 45 °C. The thermal runaway trigger temperature, defined as the cell surface temperature at the onset of thermal runaway, was recorded for all test specimens. For the pristine cells (before cycling), the thermal runaway trigger temperatures ranged from 95 °C to 125 °C, with an average of approximately 109 °C. This relatively wide range reflects variations in cell design, material quality, and manufacturing tolerances among the different battery models. For the aged cells (after 1000 cycles), the thermal runaway trigger temperatures ranged from 115 °C to 135 °C, with an average of approximately 123 °C. Notably, the aged cells exhibited consistently higher trigger temperatures compared to their pristine counterparts, with the difference ranging from 10 °C to 20 °C.
The elevation of thermal runaway trigger temperature after cycling is a counterintuitive but mechanistically explainable phenomenon. During cycling, the SEI layer on the anode surface thickens and becomes more chemically stable. This thicker SEI layer acts as a thermal barrier, delaying the onset of exothermic reactions between the lithiated anode and the electrolyte. Additionally, the loss of active lithium during aging reduces the lithium inventory in the cell, which can decrease the driving force for certain exothermic reactions. The SEI layer in aged cells contains a higher proportion of inorganic components, such as LiF and Li₂CO₃, which have higher thermal stability compared to organic components. These inorganic-rich SEI layers decompose at higher temperatures, contributing to the observed increase in thermal runaway trigger temperature.
However, it is crucial to recognize that a higher trigger temperature does not necessarily translate to improved overall safety. While the onset of thermal runaway may be delayed, the total amount of energy released during the event, the rate of heat generation, and the propagation behavior within a battery module may be different for aged cells. The aged cells have undergone significant structural and chemical changes, including potential lithium plating at the anode, electrolyte decomposition, and loss of mechanical integrity of the electrodes. These factors can influence the severity of thermal runaway events in complex ways that are not captured solely by the trigger temperature.
To establish a quantitative link between battery state of health (SOH) and thermal runaway behavior, we measured the SOH of each cell after the 1000-cycle aging test and correlated it with the thermal runaway trigger temperature. The SOH values ranged from 87.74% to 92.58%, with an average of 90.07%. The thermal runaway trigger temperatures for these aged cells ranged from 115 °C to 133 °C. When we computed the ratio of thermal runaway trigger temperature (T_TR) to SOH, a remarkable consistency emerged. The ratio T_TR / SOH fell within the narrow range of 130 to 145 for all ten aged cells, despite the variations in cell source and exact SOH value.
The detailed results for each aged cell are presented in the table below.
| Cell ID | Pre-Cycling SOH (%) | Pre-Cycling T_TR (°C) | Post-Cycling SOH (%) | Post-Cycling T_TR (°C) | Trigger Time (s) |
|---|---|---|---|---|---|
| 1 | 100 | 98.1 | 87.74 | 115 | 1215 |
| 2 | 100 | 100.1 | 88.96 | 117 | 953 |
| 3 | 100 | 98.3 | 88.61 | 116 | 1136 |
| 4 | 100 | 100.4 | 90.18 | 123 | 1028 |
| 5 | 100 | 103.7 | 90.24 | 125 | 807 |
| 6 | 100 | 103.9 | 90.25 | 125 | 762 |
| 7 | 100 | 121.0 | 90.26 | 128 | 808 |
| 8 | 100 | 125.0 | 91.26 | 131 | 658 |
| 9 | 100 | 121.0 | 90.58 | 131 | 652 |
| 10 | 100 | 124.0 | 92.58 | 133 | 703 |
The relationship between the thermal runaway trigger temperature and SOH for the aged cells can be expressed as:
$$\frac{T_{TR}}{SOH} = k$$
where k is a constant falling within the range of 130 to 145. Using the average SOH value of 90.07% and the average thermal runaway trigger temperature of 123 °C, the average value of k is approximately 136.6. This relationship provides a practical tool for assessing the thermal safety status of energy storage batteries in the field. By monitoring the SOH through routine capacity measurements, operators can estimate the current thermal runaway trigger temperature of the battery and adjust safety protocols accordingly.
We further examined the relationship between the trigger time (the time elapsed from the start of heating to the onset of thermal runaway) and the SOH. Unlike the trigger temperature, the trigger time did not exhibit a clear correlation with SOH. The trigger times for the ten aged cells ranged from 652 seconds to 1215 seconds, with no apparent dependence on the SOH value. This lack of correlation is likely due to the influence of other factors, such as the state of the electrolyte, the condition of the separator, and the presence of any localized defects within the cell, which can significantly affect the kinetics of thermal runaway initiation. These factors vary from cell to cell and are not directly captured by the global SOH metric.
The consistency of the T_TR / SOH ratio across different cell models and manufacturers suggests that this relationship is a fundamental characteristic of the LFP chemistry under the aging conditions studied. The physical basis for this relationship lies in the evolution of the SEI layer and the loss of active lithium during cycling. As the battery ages, the SEI layer thickens and becomes more thermally stable, raising the temperature at which the exothermic reactions between the anode and electrolyte are initiated. Simultaneously, the loss of active lithium reduces the lithiation degree of the anode, which also contributes to the increased thermal stability. The ratio between the trigger temperature and SOH captures the net effect of these competing mechanisms, providing a simple yet powerful parameter for safety assessment.
The practical implications of this finding for the energy storage battery industry are significant. Current safety monitoring systems for energy storage installations primarily rely on voltage, current, and temperature measurements to detect abnormal conditions. While these parameters can identify imminent failures, they do not provide predictive information about the evolving safety margin of the battery as it ages. By incorporating SOH monitoring into the battery management system, operators can track the gradual shift in thermal stability and implement proactive measures, such as adjusting the charge-discharge limits, increasing the frequency of cell balancing, or scheduling preventive maintenance, before the battery reaches a critical safety state.
For example, consider an energy storage battery with an initial SOH of 100% and an initial thermal runaway trigger temperature of 110 °C. After several years of operation, the SOH declines to 85%. Using the relationship we established, the estimated thermal runaway trigger temperature would be approximately 116 °C, representing an increase of 6 °C due to aging. While this increase may seem modest, it shifts the safety margin relative to the operating temperature range and the protection thresholds set in the battery management system. If the system is designed with a fixed thermal runaway threshold based on the pristine cell behavior, it may not adequately account for the evolving thermal characteristics of the aged battery. Our findings provide the data needed to implement adaptive safety thresholds that evolve with the battery’s SOH.
It is important to note that the relationship between SOH and thermal runaway trigger temperature was established under specific aging conditions (1000 cycles at 45 °C). The generalizability of this relationship to other aging conditions, such as different temperatures, charge-discharge rates, or depth of discharge, requires further investigation. Different aging protocols may produce different SEI compositions and degradation patterns, which could affect the thermal stability evolution in distinct ways. For instance, aging at lower temperatures with high charge rates may promote lithium plating, which introduces a highly reactive metallic lithium phase that could significantly lower the thermal runaway trigger temperature, potentially reversing the trend observed in our study. Future work should explore a wider range of aging conditions to map the complete landscape of SOH-safety relationships.
Accelerated Aging Mechanism and Its Implications for Energy Storage Battery Lifetime Prediction
The accelerated aging observed at 45 °C compared to 25 °C provides valuable insights into the dominant degradation mechanisms and their temperature dependence. The fact that the capacity fade at 45 °C is approximately twice that at 25 °C, after the initial stabilization period, indicates that the primary degradation pathway follows an Arrhenius-type behavior with an effective activation energy that can be estimated from the acceleration factor. Using the Arrhenius equation:
$$k(T) = A \exp\left(-\frac{E_a}{RT}\right)$$
where k(T) is the rate constant at temperature T, A is the pre-exponential factor, E_a is the activation energy, and R is the universal gas constant, we can estimate the activation energy corresponding to an acceleration factor of 2 between 25 °C and 45 °C. The ratio of rate constants at the two temperatures is:
$$\frac{k(45°C)}{k(25°C)} = \exp\left[-\frac{E_a}{R}\left(\frac{1}{318.15} – \frac{1}{298.15}\right)\right] = 2$$
Solving for E_a yields approximately 50 kJ/mol, which is within the range reported in the literature for SEI growth in LFP batteries. This consistency supports the interpretation that SEI growth is the dominant degradation mechanism under our testing conditions. The activation energy of 50 kJ/mol corresponds to a temperature sensitivity that is moderate compared to other degradation mechanisms. For example, transition metal dissolution and cathode degradation typically have activation energies in the range of 80 to 120 kJ/mol, while lithium plating is less temperature-sensitive with activation energies below 30 kJ/mol. The dominance of SEI growth as the primary degradation pathway at moderate temperatures is favorable for lifetime prediction because it yields predictable, linear aging behavior that can be accurately modeled.
The linear aging regime observed after the initial cycles allows for simple extrapolation to predict the remaining useful life of the energy storage battery. If the linear degradation rate at a reference temperature is known, the degradation rate at any other temperature can be estimated using the acceleration factor. This enables the development of temperature-corrected lifetime models that can account for the thermal history of the battery in field applications. For example, if a battery operates at 35 °C for 500 cycles and then at 45 °C for another 500 cycles, the equivalent degradation at the reference temperature of 25 °C can be calculated by summing the cycle counts weighted by the acceleration factors at each temperature.
The acceleration factor at any temperature T (in °C) relative to the reference temperature of 25 °C can be expressed as:
$$AF(T) = \exp\left[-\frac{E_a}{R}\left(\frac{1}{T+273.15} – \frac{1}{298.15}\right)\right]$$
Using E_a = 50 kJ/mol, the acceleration factor at 45 °C is 2.0, at 35 °C is approximately 1.4, and at 55 °C is approximately 2.8. These values provide quantitative guidance for thermal management system design. If the objective is to limit the aging acceleration factor to below 1.5, the battery temperature should be maintained below 37 °C. This can be achieved through active cooling systems, such as air conditioning or liquid cooling, or through passive thermal management, such as phase change materials or thermal insulation.
The relationship between cycle life and temperature can be further refined by considering the nonlinear effects at very high temperatures. At temperatures above 50 °C, additional degradation mechanisms, such as electrolyte decomposition and cathode structural degradation, may become significant, increasing the aging rate beyond what would be predicted by the simple Arrhenius model based on SEI growth alone. At temperatures below 15 °C, the risk of lithium plating during charging increases, which can cause rapid capacity loss and safety hazards. Therefore, the optimal temperature range for energy storage battery operation, balancing aging rate and safety, is between 20 °C and 35 °C, with the lower end of this range being preferable for long-term stability.
The consistency of the capacity fade behavior across the five battery models in our study is encouraging for the development of universal lifetime models. However, it is important to note that all models were based on the LFP chemistry and were sourced from major manufacturers with mature production processes. Cells from smaller manufacturers or based on different chemistries, such as nickel-manganese-cobalt (NMC) or lithium-manganese-oxide (LMO), may exhibit different aging behaviors and acceleration factors. Future work should extend the methodology to other chemistries to build a comprehensive database of aging parameters for different energy storage battery technologies.
Safety Assessment Framework for Aged Energy Storage Batteries
Based on our experimental findings, we propose a framework for assessing the thermal safety status of aged energy storage batteries that integrates SOH monitoring with thermal runaway risk evaluation. The framework consists of three tiers of assessment, each providing increasing levels of detail and accuracy.
The first tier is based on the relationship between SOH and thermal runaway trigger temperature established in this study. For LFP batteries, the trigger temperature T_TR can be estimated from the SOH using the formula:
$$T_{TR} = SOH \times k$$
where k is taken as the central value of the observed range, 136.6. This estimate provides a baseline for the thermal stability of the aged battery and can be used to set adaptive safety thresholds in the battery management system. For example, if the maximum allowable temperature in the system is set at T_max, and the estimated T_TR is less than T_max plus a safety margin, the system can flag the cell for closer monitoring or maintenance.
The second tier involves periodic thermal characterization of representative cells from the population. While direct thermal runaway testing is destructive and cannot be performed on all cells, a sampling approach can provide validation of the SOH-based estimates and capture any batch-to-batch variations. Cells that have reached a predefined SOH threshold, such as 80% or 70%, can be removed from service and subjected to thermal runaway testing to confirm the predicted trigger temperature and to assess the severity of the thermal runaway event in terms of heat release rate, gas generation, and fire propagation potential.
The third tier incorporates additional diagnostic parameters beyond SOH to refine the safety assessment. Electrochemical impedance spectroscopy can provide information about the SEI thickness and porosity, which influence thermal stability. Differential voltage analysis can detect the onset of lithium plating, which would significantly alter the thermal runaway characteristics. Acoustic emission monitoring can detect micro-cracking and gas generation within the cell, providing early warning of developing safety issues. By integrating these multi-modal diagnostic data into a machine learning model, a more accurate and individualized safety assessment can be achieved for each cell in the energy storage battery system.
The implementation of this framework in practice requires the integration of SOH tracking algorithms into the battery management system. Incremental capacity analysis and differential voltage analysis are two methods that can be used to estimate SOH from charge-discharge data without requiring specialized hardware. These methods analyze the shape of the voltage-capacity curve during charging or discharging and extract features that correlate with the loss of active lithium and active material. The SOH estimates from these methods have been shown to be accurate to within 2% under laboratory conditions and can be implemented in real-time with modest computational resources.
For field applications, the SOH can be updated after each full charge-discharge cycle or periodically through scheduled conditioning cycles. The thermal runaway trigger temperature estimate can then be updated accordingly, and the safety thresholds in the battery management system can be adjusted in a feedback loop. If the estimated trigger temperature approaches the operating temperature range, the system can reduce the maximum allowable state of charge, limit the charge-discharge rate, or schedule the cell for replacement. This proactive approach to safety management represents a significant advancement over traditional threshold-based systems that react only after abnormal conditions are detected.
The economic implications of this framework are substantial. By enabling the safe operation of energy storage batteries throughout their useful life, rather than retiring them prematurely based on conservative safety margins, the total cost of ownership of energy storage systems can be reduced. The ability to predict and manage safety risks also reduces the insurance premiums and liability costs associated with energy storage installations, which have been significant barriers to deployment in some markets. Furthermore, the framework provides a basis for developing second-life applications for retired energy storage batteries, where the safety assessment can determine whether a battery is suitable for less demanding applications after its primary service life.
Conclusion
Through our comprehensive investigation of accelerated aging and post-cycling safety characteristics of large-format LFP energy storage batteries, we have established several key quantitative relationships that advance the understanding of battery degradation and its implications for thermal safety. The cycling performance tests at 25 °C and 45 °C revealed that the capacity fade at the elevated temperature is approximately twice that at room temperature after the initial stabilization period, with the ratio of degradation rates converging to 2 as cycle number increases. This acceleration factor provides a quantitative basis for lifetime prediction and thermal management optimization. The relationship between the degradation rate ratio and cycle number was successfully modeled using an inverse function with a correlation coefficient of 0.9976, confirming the predictability of the aging behavior.
The thermal runaway testing of pristine and aged cells revealed that the trigger temperature for thermal runaway increases after cycling, from an average of approximately 109 °C for pristine cells to approximately 123 °C for cells aged to an average SOH of 90.07%. This increase is attributed to the evolution of the SEI layer during aging, which becomes thicker and more thermally stable. Most importantly, we discovered that the ratio of thermal runaway trigger temperature to SOH falls within a narrow range of 130 to 145 for all aged cells tested, providing a practical tool for assessing the thermal safety status of energy storage batteries based on their SOH. This relationship enables the implementation of adaptive safety management strategies that evolve with the battery’s state of health, enhancing both safety and economic viability of energy storage systems.
The findings from this study contribute to the broader goal of developing reliable, safe, and cost-effective energy storage solutions that can support the global energy transition. By providing quantitative relationships between operating conditions, aging rates, and safety characteristics, our work equips battery manufacturers, system integrators, and end-users with the knowledge needed to design, operate, and maintain energy storage systems with confidence. Future work should extend this research to other battery chemistries, explore the effects of different aging protocols on safety evolution, and develop real-time monitoring algorithms that can implement the SOH-based safety assessment framework in practical energy storage installations.