Safety is a key aspect of battery energy storage systems. Understanding safety aspects of large-size cells is complicated not only by the interplay of multiple physical phenomena but primarily due to limitations in experimental capabilities that can accurately measure all the relevant signals. Due to the limits of battery testing, information such as temperature can be only measured in a practical way. Recognizing this gap, a combined calorimetry and modeling approach is developed for large-size cells to shorten design cycles and reduce cost to optimize batteries with improved safety. Safety of large-size cells is complementarily determined by safety features implemented at multi-levels including material, chemistry, cell and system design. Calorimetry methods have been widely adopted to analyze battery failure mechanisms and hazard under electrical, mechanical and thermal abuse conditions at chemistry and cell levels [1][2]. It prefers sample spatially uniform responses to abuse conditions and is lack of capability to discern orderly sequences to characterize failure evolution within a cell and failure propagation among cells. In this presentation, an integrated multiphysics li-ion battery safety model built on a nonlinear multi-scale multi-domain (MSMD) model framework [3] is introduced. Modularized hierarchical architecture of the MSMD framework allows independent choice of submodels. In the combined approach, the models are adept at utilizing experimental results. calorimetry results serve as guideline to tune the integrated multiphysics model, such as determining physical behaviors being simulated in the submodels and identifying corresponding modeling parameters. Modeling results are then correlated with experimental data to identify potential experimental uncertainties. Safety performance at higher levels can be directly estimated using this model without additional model calibration. The combined approach enables battery safety being addressed efficiently across multiple cells at real-world scenario. It provides insights on functionalities of safety mechanisms and its contribution to overall battery safety. The approach has been applied to study transient gas flow behaviors within an 18650 battery cell subject to internal short circuit and validate safety technical assumptions for the design of a novel packaging architecture for lithium-ion batteries [4]. Moreover, this approach can quantify non-measurable properties and its effects on battery failure severity, such as thermal interfacial properties among large-size cells. Acknowledgement This study was supported by Computer Aided Engineering for Batteries (CAEBAT) project of the Vehicle Technologies Office, Office of Energy Efficiency and Renewable Energy, U.S. Department of Energy. The research was performed using computational resources sponsored by the Department of Energy's Office of Energy Efficiency and Renewable Energy, located at the National Renewable Energy Laboratory. Reference [1] C. Lampe-Onnerud, et al., Safety Studies on Lithium-Ion Batteries by Accelerating Rate Calorimetry, Battery Conference on Applications and Advances, 1999 [2] W. Walker, et al., Statistical characterization of 18650-format lithium-ion cell thermal runaway energy distributions, NASA Aerospace Battery Workshop, Huntsville, Alabama, 2017 [3] G.-H. Kim, K. Smith, K.-J. Lee, S. Santhanagopalan, and A. Pesaran, J. Electrochem. Soc., 58, 8 (2011) [4] Cadenza Innovation, INC (2018). ARPA-E Final Technical Report, Novel low cost and safe Lithium-Ion Electric Vehicle Battery, DE-AR0000392