Abstract
A 2D filter is proposed for battery incremental capacity curve filtering in a cyclic aging test. The filter works in two directions, namely, from time to time and from batch (cycle) to batch. In details, a simple low-pass filter is applied in the batch direction, and a bias-corrected Gaussian filter is applied in the time direction. Experimental results show that the root-meansquare-error of the proposed method is 20% lower than the neural-network-based benchmarking algorithm. Over-fitting and under-fitting could also be resolved.
Keywords 2D filter, incremental capacity analysis (ICA), Gaussian filter
Copyright ©
Energy Proceedings