Open AccessKalman filter applied to Thevenin’s modeling of a lead-acid battery
Author(s) -
Jose Alfredo Palacio-Fernádez,
Edwin García Quintero
Publication year - 2022
Publication title -
international journal of power electronics and drive systems/international journal of electrical and computer engineering
Language(s) - English
Resource type - Journals
eISSN - 2722-2578
pISSN - 2722-256X
DOI - 10.11591/ijece.v12i2.pp1350-1357
Subject(s) - lead–acid battery , kalman filter , battery (electricity) , thévenin's theorem , signal (programming language) , noise (video) , state of charge , computer science , matlab , equivalent circuit , control theory (sociology) , filter (signal processing) , voltage , extended kalman filter , state of health , electrical engineering , engineering , physics , artificial intelligence , power (physics) , control (management) , quantum mechanics , image (mathematics) , computer vision , programming language , operating system
This article determines the internal parameters of a battery analyzed from its circuit equivalent, reviewing important information that can help to identify the battery’s state of charge (SOC) and its state of health (SOH). Although models that allow the dynamics of different types of batteries to be identified have been developed, few have defined the lead-acid battery model from the analysis of a filtered signal by applying a Kalman filter, particularly taking into account the measurement of noise not just at signal output but also at its input (this is a novelty raised from the experimental). This study proposes a model for lead-acid batteries using tools such as MATLAB ® and Simulink ® . First, a method of filtering the input and output signal is presented, and then a method for identifying parameters from 29 charge states is used for a lead-acid battery. Different SOCs are related to different values of open circuit voltage (OCV). Ultimately, improvements in model estimation are shown using a filter that considers system and sensor noise since the modeled and filtered signal is closer to the original signal than the unfiltered modeled signal.