Search

A pack of batteries is one of the most important and expensive assemblies for an electric vehicle. A pack of batteries is comprised of several batteries connected in series. The number of the batteries connected depends on the operating voltage of the vehicle's on-board system as well as on the individual characteristics of the batteries used, e.g. the operating voltage of a single cell. One or several cells of a pack of batteries could be damaged if improperly exploiting an electric vehicle- excessively discharging the batteries or overloading the electric vehicle. If a self-converted vehicle does not use an intellectual BMS (battery management system) that can identify and register voltage drop for any individual cell in the high-load regime, e.g. when accelerating, it is difficult to identify and change the cells damaged. In case a cell does not demonstrate a complete failure, it is almost impossible to identify a defect in any regime other than the load regime. The research developed and compared three different methods for identifying defective battery cells. The methods were approbated on a converted Renault Clio. The experiment involved making voltage measurements in road tests, running the electric vehicle on a roll test bench and making voltage measurements of maximally discharged batteries in the no-load regime. A comparison of the measurement results revealed that the measurements made in the road tests were the most accurate and useful. After the experiment, the defective battery cells were replaced, thereby restoring the performance of the battery pack. [ABSTRACT FROM AUTHOR]

The studies on an integrated approach for the battery (cell level), battery pack (system level) and battery pack enclosure (system level) are nascent till date. The present work proposes a combined approach for studying the coupled electrochemical-mechanical performance of the battery (cell level), battery pack and enclosure (system level). In this context, five research problems are undertaken involving the design parameters for the performance evaluation of batteries. The first issue undertaken is battery performance under stress conditions and for that varying initial stress and real-time stresses were applied, and variation in the capacity of the battery packs was studied. Initial stress was found to be in negative relation with the battery capacity. For understanding the mechanical damage (leakage, deformation, etc.), characterization of the battery cell under impact loading was done by subjecting battery cells to different impact loads and impact analysis was done using six statistical features extracted from the impact force processed signal of each test, namely: peak value, mean value, variance, skewness, kurtosis and pulse width. The third issue regards the manufacturing defects in the cell, which can potentially lead to inconsistency of cells performance in a pack and thus lower its total cycle life. Support vector classification clustering methods were applied to form a battery pack with similar performing cells and performance of the pack was compared with industrial packs in which earlier one was found to be better. The fourth research challenge involved the mechanical performance parameters such as strength, deformation, and natural frequency, which, in addition to temperature, are also critical to battery pack safety in the event of impact and vibrations under driving conditions. ANSIS and NSGA-II based approaches were used to optimize the battery design and resultant device shown maximum deformation reduced by 22.2%, the value of minimum inherent frequency improved by 3.2% and the value of mass decreased by 11.6%. The fifth research problem highlighted how, due to the enormous production of electric vehicles, by 2020, there will be 250,000 tons of batteries that need to be recycled. A combined experimental-numerical framework has been thereby suggested to mitigate it. The study resulted in the formation of an online monitoring system and safety design of batteries at both cell and system level (pack and enclosure). [ABSTRACT FROM AUTHOR]