The development of a battery management system with special focus on capacity estimation and thermal management

Lithium ion batteries are instrumental in tackling the challenges of global warming. They have shown great utility in electric and hybrid vehicles. However, challenges with regard to performance and safety such as capacity fade and thermal runaway need to be accounted for in the implementation of these battery systems. One way is through battery management systems that monitor and control various aspects of the battery’s operation. At the heart of the battery management system is an analytical model of the battery. This thesis proposes a battery management system which uses a “lowuses a “low-order” physics- based battery model that estimates capacity and optimally manages the temperature of the battery. A capacity estimation methodology is proposed that uses the state of charge estimate from an extended kalman filter and the inverse of the coulomb counting equation to estimates the “instant” capacity of the battery. This instant value is then used in an averaging calculation that uses saturation limits and a time delay to obtain a value for the capacity that is representative of the battery. This value is then feedback into the kalman filter. The capacity estimate obtained through this method was between 2 and 8 % off of the true value. A thermal management system is also proposed that optimally controls a fan to cool a lithium ion battery. The thermal management system was developed and tested in a simulated environment. First, the fan model was integrated with the battery model and simulations were run to test the open loop temperature response of the battery to the fan cooling while varying the input voltage of the fan the current demanded of the battery. From this data an operating point was chosen, the system was linearized, and a linear quadratic controller was designed and implemented. The controller was sluggish when faced with a temperature perturbation in the absence of a current demand increase but drove the temperature change to zero. In the presence of a current change controller drove the state to a nonzero steady state value. The same result occurred when a disturbance rejection mechanism was applied to the controller.