A high safety standard of battery systems is essential for the operation of Battery Electric Vehicles. Especially for future battery systems with high energy densities, the consideration of Thermal Runaway (TR) and its Propagation (TP) is of great importance in order to further ensure high safety standards within automotive applications. However, the severity of a Thermal Runaway event and its varying impact on neighboring cells makes it challenging to identify optimal Thermal Propagation preventing measures. The use of model-based TR prediction based on experimental thermal analysis of cell components is one key factor to further understand the variability of a TR event and hence optimize battery safety designs. This study examines decoupled heat generation of cell components and their impact to the overall heat generation during TR under special consideration of their heat output variability. Therefore, differential scanning calorimetry (DSC) tests on individual cell components and their mixtures with various repetitions of a state of the art automotive battery cell are carried out. With this method, the dominant exothermic reactions of the TR event can be determined. By coupling of the identified exothermic reactions, the TR behavior and its statistical scattering can be analyzed. The results on cell level can then be used to evaluate the influence of TR severity on battery system level quantities. This talk will provide an overview of the thermal analysis results on cell component level using statistical methods as well as an outlook for their use in predictive TP modeling based on varying TR behavior.