Study on standard temperature-rise curve for battery energy storage fire based on multi-condition data fusion

The thermal runaway fire of lithium-ion battery energy storage systems shows nonlinear propagation and extreme thermal shock. Traditional standard temperature-rise curves, such as ISO 834 and HC (Hydrocarbon), cannot accurately characterize the thermal environment evolution of battery energy storage system fires. Based on full-scale combustion tests of 30 and 220 kWh multi-scale energy storage cabinets, this study proposes a method combining time-dimensionless normalization, principal component analysis–KMeans clustering, worst-case envelope extraction, double-exponential fitting, time-axis calibration, and dynamic-weight fusion. It converts multi-condition, multi-point temperature data into a unified standard curve. Results show that top-interior measuring points have strong thermal accumulation and stable features for representing the worst-case fire scenario. The derived battery energy storage system’s curve heats faster than ISO 834 but slower than HC. This method removes test-condition deviations and offers a robust engineering benchmark for battery energy storage system’s fire-resistance design, material selection, fire protection evaluation, and safety assessment.