Haussmann, P. and Melbert, J., "Internal Cell Temperature Measurement and Thermal Modeling of Lithium Ion Cells for Automotive Applications by Means of Electrochemical Impedance Spectroscopy," SAE Int. J. Alt. Power. 6(2):2017, doi:10.4271/2017-01-1215.
Battery safety is the most critical requirement for the energy storage systems in hybrid and electric vehicles. The allowable battery temperature is limited with respect to the battery chemistry in order to avoid the risk of thermal runaway. Battery temperature monitoring is already implemented in electric vehicles, however only cell surface temperature can be measured at reasonable cost using conventional sensors. The internal cell temperature may exceed the surface temperature significantly at high current due to the finite internal electrical and thermal cell resistance. In this work, a novel approach for internal cell temperature measurement is proposed applying on board impedance spectroscopy. The method considers the temperature coefficient of the complex internal cell impedance. It can be observed by current and voltage measurements as usually performed by standard battery management systems. The relevant frequency range considered for temperature measurements is chosen for high sensitivity and robust behavior and takes state of charge variations as well as aging effects into account. Transient temperature variations caused by various load profiles are analyzed in order to characterize the static and dynamic thermal properties of the cell. The resulting thermal equivalent model describes temperature changes inside the cell dependent on load current and ambient temperature. The temperature measurement approach and the thermal model are suitable for on board implementation in battery management systems. A dedicated battery excitation is not required, as signal components in the relevant frequency range are inherently present in typical driving current profiles. Thereby, significant improvements in terms of on-board diagnostics and battery safety can be achieved without any additional hardware effort.