|dc.description.abstract||In order to overcome range anxiety of electric vehicles (EVs) powered by lithium-ion batteries, more batteries are installed to increase battery capacity, which results in a long charging time that is one of the most challenging issues for EVs to be commercialized. The charging time can be reduced simply by increasing current rates, which can adversely accelerate battery degradation. Especially at low temperatures, fast charging leads to rapid degradation due to reduced ionic diffusivity and increased charge transfer resistance. Therefore, design of a charging strategy considering operating temperature for lithium-ion batteries is one of the most important issues that affect the overall performance of the batteries.
In the first part of the dissertation, a charging algorithm considering degradation is designed for room temperature. For the algorithm, an accurate reduced-order electrochemical model (ROM) considering side reaction, lithium plating, and lithium stripping is developed and experimentally validated. Afterward, the charging algorithm is designed, and the cycling results by the algorithm are further compared with those by normal charging methods, which has shown that the battery can be charged faster with less degradation using the proposed charging algorithm. However, this charging algorithm does not consider any effect of the battery heat generation and temperature.
Accordingly, a thermal model based on the ROM is developed as the second part of this dissertation. For the accurate estimation of the battery heat generation rate during operations, the detailed heat source terms considering internal processes inside the battery are mathematically formulated. The model is further validated with respect to measured heat generation rate profiles under a wide range of current rates and temperatures. In addition, further detailed analysis for heat source terms, including parameter sensitivity analysis, is conducted. The analysis results have shown that the parameters have different tendencies of sensitivity because each heat source term is affected by different parameters. Accordingly, the parameters are clustered into their sensitive SOC ranges, and as a result, a new parameter identification procedure is developed based on a three-stage stepwise identification procedure. The identified parameters are further validated, which has shown that the proposed method is able to estimate the parameters accurately with less experiment set.
In the third part, the charging algorithm is extended by considering subzero temperatures, where mechanical degradation is also considered. The model is further experimentally validated at subzero temperature, showing that a steep concentration gradient affects the mechanical degradation at the cathode electrode. Finally, a charging algorithm at subzero temperature is designed, where nonlinear model predictive control and genetic algorithm are applied to optimize charging profile and start-up heating strategy. The cycling results by the algorithm are further compared with those by normal charging methods, which has shown that the proposed algorithm is able to charge the battery faster with higher charging capacity than those using normal charging methods. In addition, the degradation rate by the proposed algorithm is comparable to the normal charging methods.
As a closing work to the dissertation, we further propose an implementation of various optimization techniques that can be applied to the ROM. Since the current optimization techniques still have some drawbacks such as long calculation time or additional tasks of finding optimal weighting factor for a trade-off between different objective functions, a comparison of various online or offline optimization techniques may bring a solution for future direction for the optimization of the parameter identification or charging algorithm under a wide range of temperatures.||en_US