An accurate algorithm for lithium polymer battery SOC estimation is proposed based on adaptive unscented Kalman filters (AUKF) and least square support vector machines (LSSVM). A novel approach using the moving window method is applied, with AUKF and LSSVM to accurately establish the battery model with limited initial training samples. The effectiveness of the moving window modeling method is validated by both simulations and lithium polymer battery experimental results. The measurement equation of proposed AUKF method is established by the LSSVM battery model, and AUKF has the advantage of adaptively adjusting noise covariance during the estimation process.
In addition, the developed LSSVM model is continuously updated online with new samples during the battery operation, in order to minimize the influence of the changes in battery internal characteristics on modeling accuracy and estimation results after a period of operation. Finally, a comparison of accuracy and performance between AUKF and UKF is made. Simulation and experiment results indicate that the proposed algorithm is capable of predicting lithium battery SOC with a limited number of initial training samples.Index Terms-Lithium polymer battery, moving window method, modeling, least square support vector machine (LSSVM), adaptive unscented Kalman filter (AUKF), state of charge (SOC). Manuscript
is an open access repository that collects the work of Arts et Métiers ParisTech researchers and makes it freely available over the web where possible.
AbstractThis paper presents a 2-D real-time modeling approach for a proton-exchange-membrane fuel cell (PEMFC). The proposed model covers multi-physical domains for both fluidic and electrochemical features, which considers in particular the flow field geometric form of fuel cell. The characteristics of reactant gas convection in the serpentine gas pipeline and diffusion phenomenon through the gas diffusion layer (GDL) are thoroughly considered in fluidic domain model. In addition, a three levels iterative solver is developed in order to accurately calculate the implicit spatial physical quantities distribution in electrochemical domain. Moreover, the proposed 2-D real-time modeling approach uses a numerical method to achieve a fast execution time, and can thus be further easily applied to any realtime control implementation or online diagnostic system. After experimental validation under different fuel cell operating conditions, an iterative Least Angle Regression (LAR) method is used to efficiently and accurately perform the global parameters sensitivity analysis based on Sobol definition. The online analysis results give an insight into the influences of modeling parameters on fuel cell performance.The effect of interactions between parameters' sensitivities is especially investigated, which can provide useful information for degradation understanding, parameters tuning, re-calibration of the parameters and online prognostic.Keywords: Proton exchange membrane fuel cell, flow field geometric form, global parameters sensitivity, effect of interactions.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.