Influence of Uncertainty of Thermal Conductivity on Prediction Accuracy of Thermal Model of Lithium-Ion Battery

This study employed the transient plane source (TPS) method to measure the battery's thermal conductivity. The probe heated the battery and collected its temperature. Based on the measured temperature, the thermal conductivity was calculated. Then, this tested thermal conductivity is compared with the theoretical value to get the prediction error of the theoretical algorithm. For the 27 Ah battery, the relative error of thermal conductivity through the material layer k(x) is 30.2%, while those of thermal conductivities along material layers k(y) and k(z) are 89.8%. Then, a 3-D thermal model based on the thermal network was established, and it applied the calculated and measured thermal conductivity to quickly predict battery temperature distribution at discharging rate from 1 degrees C to 6 degrees C. According to the results, the theoretical model for thermal conductivity should be used at a discharging rate below 3 degrees C, or a great prediction error is produced. To further improve the prediction accuracy of battery temperature field at high discharging rates, the error set of the thermal conductivity was built, and the threshold of the error was explored. The relative error of k(x) should vary between 15% and -15% and those of k(y) and k(z) should be below -45%. Moreover, the prediction error of the battery temperature is small, as the relative errors of k(y) and k(z) increase from 15% to 90%.