A New Time Frequency Technique Of Charging And Discharging Signal Analysis For Battery Parameters Identification System

Selamat, Nur Asmiza and Abdullah, Rahim and Ranom, Rahifa and Shamsudin, Nur Hazahsha and Karis, Mohd Safirin (2018) A New Time Frequency Technique Of Charging And Discharging Signal Analysis For Battery Parameters Identification System. Project Report. UTeM, Melaka, Malaysia. (Submitted)

Text A New Time Frequency Technique Of Charging And Discharging Signal Analysis For Battery Parameters Identification System.pdf - Submitted Version Restricted to Registered users only Download (352kB)

Abstract

Battery is an alternative option for future energy demand. Numerous type of battery is offer on the market to propel portable power and its makes the task of selecting the right battery type is crucial. In addition, battery lifetime degrades and directly affects by load performance. Hence, reliability and safety operation of the battery not guarantee. In return, safety precautions by monitoring battery performance from charging/discharging signals behavior suggested. Analyse the battery charging/discharging signals become challenging as the signal characteristic appears at very low frequency. Therefore, fast and accurate analysis in estimating battery parameters for real-time monitoring system should be proposed and developed. This research presents analysis of the battery charging/discharging signals using a spectral analysis technique, namely periodogram and time-frequency distributions (TFDs) which are spectrogram and S-transform techniques. The analysed batteries are lead acid (LA), nickel-metal hydride (Ni-MH) and lithium-ion (Li-ion). From the equivalent circuit model (ECM) simulated using MATLAB, constant charging/discharging signals are presented, jointly, in time-frequency representation (TFR). From the TFR, battery signal characteristics are determined from the estimated parameters of instantaneous of total voltage (VTOT (t)), instantaneous of average voltage (VAVG (t)) and instantaneous of ripple factor voltage (VRF (t)). Hence, an equation for battery remaining capacity as a function of estimated parameter of VRF (t) using curve fitting tool presented. In developing a real-time automated battery parameters estimation system, best TFD chosen in terms of accuracy of battery parameters, computational complexity in signal processing and memory size. Advantages in high accuracy for battery parameters estimation and low in memory size requirement makes S-transform technique selected to be the best TFD. The accuracy of the system verified with parameters estimation using ECM for each type of battery at a different capacity. The field-testing results show that average mean absolute percentage error (MAPE) is around four percent. Thus, implementation of S-transform technique for real-time automated battery parameters estimation system is very appropriate for battery signal analysis. At the end of this research, an automated system for real-time battery parameters identification using a new time frequency technique designed and develops then applied in green technology industries.

Actions (login required)

View Item