| Summary: | Accurate on-board temperature monitoring of lithium-ion batteries is important for safety and control purposes. Impedance temperature detection (ITD) is a promising approach for temperature estimation, whereby the internal cell temperature is directly inferred from online electrochemical impedance spectroscopy (EIS) measurements at a single frequency. Previously, ITD was used to infer the volume-average cell temperature; the present work focuses on extending ITD to enable estimation of the spatially-resolved temperature distribution of cells with internal temperature gradients. Two novel hybrid methods for temperature monitoring are introduced, based on combining impedance measurements with (i) an additional surface temperature measurement, and (ii) a thermal model. These methods predict the temperature distribution of the cell in either 1-D or 2-D, and can therefore identify localised hot spots, and hence the global maximum cell temperature. In each case, the methods are experimentally validated using cylindrical LiFePO4 cells (26650 for the 1-D experiments, 32113 for the 2-D experiments) monitored with periodic 215 Hz impedance measurements, and fitted with an internal thermocouple and one or more surface thermocouples for validation. Method (i) is shown to be more accurate than a standard ITD method based on impedance measurement only: e = 0.6?C vs. 2.6?C respectively, over a 3500 s drive cycle. In method (ii), the impedance measurement forms part of a state/parameter estimation algorithm; in this case, the performance of an extended Kalman filter using impedance measurement is shown to be comparable - although slightly inferior - to an equivalent Kalman Filter using a conventional surface temperature measurement. This work also presents a novel low-order 2-D thermal model based on the spectral-Galerkin (SG) method. The model can be used in conjunction with the proposed hybrid methods or in a conventional temperature monitoring scheme. Time- and frequency domain simulations show that the SG model using as few as 4 states is capable of accurately modelling the thermal dynamics of a large format cylindrical cell with a highly transient heat generation input. The model can account for different external temperatures and/or convection coefficients at each surface - a generality which makes it suitable for simulating various battery cooling configurations.
|
|---|
