Abstract-The voltage of lithium ion batteries is usually monitored to prevent overcharge and overdischarge. For battery packs consisting of hundreds of cells, monitoring the voltage of every single cell adds significant cost and complexity to the battery management system (BMS). Reducing voltage sensing by only measuring the total voltage of multiple cells in series connection is desirable if the state of charge (SOC) of individual cells can be correctly estimated. Such goal cannot be achieved by an extended Kalman filter, because the cell SOCs are not observable in the linearized battery string model. In this paper, an observer based on solving simultaneously multiple nonlinear equations along the trajectory of SOC evolution is used for the estimation problem. Existence of the solution depends on the nonlinearity of the battery voltage-SOC relationship. The observer is applied to a LiFePO 4 /graphite battery string with 2 cells, where the individual cell SOCs are observable in low and high SOC ranges. Experimental results show good convergence of SOC and voltage estimation, indicating that this new methodology can be applied to, at least, halve the voltage sensing in a battery pack.

Efficient and safe battery charge control is an important pre-requisite for large-scale deployment of clean energy systems. This paper proposes an innovative approach to devising op-timally health-conscious fast-safe charge protocols. A multi-objective optimal control problem is mathematically formulated via a coupled electro-thermal-aging battery model, where elec-trical and aging sub-models depend upon the core tempera-ture captured by a two-state thermal sub-model. The Legendre-Gauss-Radau (LGR) pseudo-spectral method with adaptive multi-mesh-interval collocation is employed to solve the result-ing highly nonlinear six-state optimal control problem. Charge time and health degradation are therefore optimally traded off, subject to both electrical and thermal constraints. Minimum-time, minimum-aging, and balanced charge scenarios are exam-ined in detail. The implications of the upper voltage bound, am-bient temperature, and cooling convection resistance to the opti-mization outcome are investigated as well. 1