Range Anxiety Alleviation for Electric Vehicles using a Dual Extended Kalman Filter with a Nonlinear Battery Model
MetadataShow full item record
The purpose of this thesis was to find a solution to combat range anxiety in electric vehicles (EV). Range anxiety was defined as the fear that the driver will not reach their destination and is a major barrier to increase EV uptake worldwide. Parameter and state of charge (SoC) estimation for the Revolve NTNU battery cell were done using an enhanced self-correcting (ESC) battery model in combination with a dual Extended Kalman Filter (EKF). Next, a method was developed to extract high resolution elevation data from a given Google Maps route. This elevation information was then attached to three different speed profiles and passed as an argument into an algorithm which calculated a SoC estimate. The algorithm used vehicle modelling equations along with the dual EKF to estimate SoC, as well as extrapolated range and time to go until battery pack depletion. Range results comparing the simulation to a real Chevy Volt using the same speed profiles showed a 23 % error. The validation of the range algorithm required a different battery cell, which made it impossible to use the dual EKF tuned to the Revolve cell for SoC estimation. Instead, coulomb counting (CC) was implemented in SoC estimation for the validation section. In addition, tests using three different elevation profiles concluded that topography had a major influence on extrapolated range. Lastly, different situations were observed where the EKF had superior SoC estimation performance compared with CC. More focus on speed profile generation will enable future EV users to use their GPS to estimate range more accurately and thereby help alleviate range anxiety.