Real-time monitoring of temperature distribution in a lithium-ion battery pack

Abstract

Intelligent battery management systems aim to continuously diagnose the state-of-health (SOH) of cells in a battery pack and predict performance, reliability and safety issues. Cells with a higher internal resistance component will generate more heat and can be identified by the battery management system (BMS) from this thermal behavior. Placing a temperature sensor on each cell in a battery pack containing thousands of individual units might not be feasible due to practical limitations or budgetary constraints. As such, this paper describes an approach to real-time monitoring of a pack's temperature distribution using a small, configurable number of temperature sensors with the aid of interpolation and extrapolation algorithms. Of the five algorithms initially considered, only the nearest neighbor and inverse distance weighting algorithms were implemented and compared in terms of accuracy and aesthetic practicality. The temperature distributions are displayed using a 3D visualization framework inside an option-rich user interface. The effect the number of temperature sensors and sample resolution had on data accuracy and latency were also investigated. With accurate and timeous data describing the behavior of the cells in a LIB pack, the BMS can better achieve its purpose in terms of monitoring, diagnosis and control