ENERGY ESTIMATION OF NCA BATTERY BY CONSIDERING THE UNBALANCED CELLS AND BATTERY WORKING AREA

• Main Author: Edison (NIM : 23315012), Frans
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/22203
• Institution: Institut Teknologi Bandung

Electric vehicles begin to replace conventional fossil-fueled vehicles. Indonesian government begin to encourage the conversions to electric vehicles (EV). The most common source of energy of electric vehicles until this daya is the battery, especially on small scale electric vehicles like e-bikes. As the main energy source, the battery should always be in good performance so the EV can operate optimally. One information that required by EV users is the remaining distance that the EV can be driven. To obtain this information, it is necessary to measure the remaining energy contained in the battery. <br /> <br /> This study aims to estimate the remaining energy of the battery in series arrangement while taking into account the imbalance between cells and the work area of the battery. The method that used are energy counting and support vector regression. Energy conting are performed by accumulating power coming out of the battery when discharging or entering the battery during charging. The power is accumulated over time and compared to its nominal energy to obtain state of energy (SOE) of the battery. This method requires an initial value as a reference for calculation for the next iteration. In the first stage, the energy counting are performed on a single cell, which is cycled with three variations of current rates from the initial state 100% fully charged, to obtain individual cell characteristics which unaffected by other cells that occur on cells arranged in series. The data from one cell is arangged into lookup table that expresses the relation between voltage, SOE, and current. The Lookup tables are used as training data for Support Vector Regression (SVR) to generate models for estimating cell’ SOE. <br /> <br /> Electric vehicles begin to replace conventional fossil-fueled vehicles. Indonesian government begin to encourage the conversions to electric vehicles (EV). The most common source of energy of electric vehicles until this daya is the battery, especially on small scale electric vehicles like e-bikes. As the main energy source, the battery should always be in good performance so the EV can operate optimally. One information that required by EV users is the remaining distance that the EV can be driven. To obtain this information, it is necessary to measure the remaining energy contained in the battery. <br /> <br /> This study aims to estimate the remaining energy of the battery in series arrangement while taking into account the imbalance between cells and the work area of the battery. The method that used are energy counting and support vector regression. Energy conting are performed by accumulating power coming out of the battery when discharging or entering the battery during charging. The power is accumulated over time and compared to its nominal energy to obtain state of energy (SOE) of the battery. This method requires an initial value as a reference for calculation for the next iteration. In the first stage, the energy counting are performed on a single cell, which is cycled with three variations of current rates from the initial state 100% fully charged, to obtain individual cell characteristics which unaffected by other cells that occur on cells arranged in series. The data from one cell is arangged into lookup table that expresses the relation between voltage, SOE, and current. The Lookup tables are used as training data for Support Vector Regression (SVR) to generate models for estimating cell’ SOE.