Lithium ion battery modeling, estimation, and aging for hybrid electric vehicle applications.

摘要

Reducing greenhouse gas emissions and improving the fuel efficiency of automobiles, trucks, and buses can be achieved by partial and full electrification of the vehicle sector. Lithium ion battery technology is the leading candidate for vehicle electrification. Despite many advantages of lithium ion battery technology, over-conservative pack design due to a lack of advanced battery management hinders its widespread deployment in the transportation sector. This dissertation introduces a model-based approach for safe and efficient advanced lithium ion battery management.;Low order, explicit models of lithium ion cells are critical for real-time battery management system (BMS) applications. Li-ion cell response varies significantly with temperature and cell temperature measurements are often available. This study presents a 7th order, single particle model with electrolyte diffusion and temperature dependent parameters (ESPM-T model). The impedance transfer function coefficients are explicit in terms of the model parameters, simplifying the implementation of temperature dependence yet providing an accurate model. The 7 th order, linear, electrolyte enhanced, single particle model (ESPM) is used as the basis for a Luenberger SOC observer for a lithium ion cell. Isothermal and non-isothermal observer performances are compared with a commercially-available finite volume code and the benefits of temperature measurement are shown for a wide range of temperature and pulse C-rates.;The ESPM is then extended to a nonlinear, electrolyte-enhanced, single particle model (NESPM), which includes nonlinearities associated with open circuit voltage and Butler-Volmer (B-V) kinetics. The model is validated with experimental full charge, discharge, and HEV cycles from 4.5 Ah high power and 20 Ah high energy graphite (gr)/LiFePO4 (LFP) cells. The NESPM is capable of operating up to 3C constant charge-discharge cycles and up to 25C and 10 sec charge-discharge pulses within 35-65% state of charge (SOC) with less than 2% error for the 4.5 Ah high power cell. For the 20 Ah high energy cell, the NESPM model is capable of operating up to 2 C constant charge-discharge cycles and up to 10C and 10 sec charge-discharge pulses within 30-90% SOC window with 3.7% maximum error.;An aging model due to solid electrolyte interphase layer growth is added to the NESPM model. The NESPM aging model is then simplified to obtain explicit formulas for capacity fade and impedance rise that depend on the battery parameters and current input history. These simple aging models can be implemented in online model based battery SOH estimation. The formulas show that aging increases with SOC, operating temperature, time, and root mean square (RMS) current. The formula predicts that HEV current profiles with the (i) same average SOC, (ii) small SOC swing, (iii) same operating temperature, (iv) same cycle length, and (v) same RMS current, will have the same cell capacity fade.;The single cell ESPM-T model is extended to a pack model with three cells in parallel to develop thermal management strategies to extend battery life within a desired performance window. Instead of defining battery End of Life (EOL) as an arbitrary percent of capacity loss, it is defined as the cycle number when.