An adaptive battery monitoring system for an electric vehicle

Abstract

In electric vehicles it is important to know the state of charge of the batteries in order to prevent vehicle strandings and to ensure that the full range of the vehicle is exploited. It is also useful to know state of health information about the batteries in the battery bank, This information can be used to predict when the batteries need replacing and can also identify batteries that are not performing optimally within the battery bank.This thesis describes a battery monitoring system that is able to calculate the state of charge and state of health of niultiple batteries in a battery bank, It has been designed specifically to monitor lead-acid batteries in an electric car environment using noninvasive measurement techniques. The monitor incorporates an adaptive monitoring method, which is based on coulometric measurements when the batteries are under load and predicted open circuit voltage measurements under no-load conditions.The battery monitor is micro controller based and uses remote battery monitoring modules to make the necessary battery measurements. Information is presented to the user of the car in the form of a state of charge meter on the instrument panel, similar to a fuel gauge in a conventionally power vehicle, and an alphanumeric LCn panel on the car's dashboard. . Aspects of both the monitor hardware and software are considered in this thesis.Results obtained from bench tests of the monitor are presented which are followed by an evaluation of the monitor's performance. Consideration is also given to possible future improvements to the monitoring system. AcknowledgmentsThere are a number of people who have assisted me over the duration of this project. Firstly thanks should go to my supervisor Dr Richard Duke for his patience, guidance and advice throughout my project. Dr Simon Round also deserves thanks for his advice and work attempting to get the electric car operational.My thanks go also to Ron Battersby and Ken Smart for their technical assistance and work on the electric car. I would also like to thank Dermot Sallis for manufacturing the many PCBs for this project and Warwick Earl for his assistance with equipment for the low temperature battery testing.In would like to thank the members of power electronics postgraduate group, particularly Dave Ingram, Arnold Memelink, Jan Wiik, Adam Taylor and Hamish Laird, for their good advice and for providing a good working environment in the lab.

Battery health determination by subspace parameter estimation and sliding mode control for an all-electric Personal Rapid Transit vehicle — the ULTra

“…This degradation can be approximated with the linear relationship described in (7) (7) where N is the test number.…”

“…Whilst SoC characterisation techniques have been published for lead-acid battery technologies [5][6][7][8][9], SoH determination is less well-reported, primarily because the SoH is a qualitative rather than quantitative measure of the ageing of a battery [3,4] and its ability to provide its nominal capacity over extended lengths of time. In most cases, the SoH is estimated by monitoring the degradation of specific battery model parameters, as suggested in [5], or close inspection of operational phenomena such as the Coup de Fouet technique used in [6].…”

“…Many authors have investigated modelling strategies for accurate SoC estimation, such as coulomb-counting methods [7], electrochemical battery models [6], or equivalent circuit battery models [5,8,9,10].…”

Novel battery model of an all-electric personal rapid transit vehicle to determine state-of-health through subspace parameter estimation and a Kalman Estimator

“…2. ULTra driving cycle current-profile Whilst SoC characterisation techniques have been published for lead-acid battery technologies [2][3][4][5][6], SoH determination is less well-reported, primarily because the SoH is a qualitative rather than quantitative measure of the ageing of a battery [7,8]. In most cases, the SoH is estimated by monitoring the degradation of specific battery model parameters, as suggested in [2], where SoH deterioration is demonstrated through use of the wellknown Randles' battery circuit model [9] using static parameters in a closed-loop Extended Kalman Filter (EKF).…”

Multiphysical Lithium-Based Battery Model for Use in State-of-Charge Determination

“…SOC estimation algorithms, based on combined current and voltage measurements such as the algorithm presented in this paper, are thus implemented on battery management systems (BMSs). This helps improve battery life duration [5]- [7] and ensure their safe operation, which could, for example, be threatened if the SOC went under or above the given limits.…”