Batteries, Vol. seven, Pages 35: Unified Evaluation Framework for Stochastic Methods Placed on Remaining Useful Existence Prognosis Problems
Electric batteries doi: 10. 3390/batteries7020035
Authors: Mikel Arrinda Mikel Oyarbide Haritz Macicior Eñaut Muxika
A unified evaluation system for stochastic tools will be developed in this document. Firstly, we provide a set of already existing quantitative and qualitative metrics that rate the relevant factors of the performance of a stochastic prognosis formula. Secondly, we provide innovative guidelines to detect plus minimize the effect associated with side aspects that socialize on the algorithms’ functionality. Those aspects are associated with the input uncertainty (the uncertainty on the particular data and the earlier knowledge), the parametrization technique and the uncertainty distribution method. The proposed assessment framework is contextualized upon a Lithium-ion battery Left over Useful Life prognosis issue. As an example, the Particle Filter is evaluated. On this example, two different data sets taken from NCA aged batteries and two semi-empirical aging models available in the literature fed up the Particle Filter under evaluation. The obtained results present that the proposed construction gives enough details in order to take decisions about the particular viability of the chosen algorithm.