LITHIUM - ION BATTERY AGED PREDICTION BASED ON CAPACITY USING MACHINE LEARNING

• Main Author: Mulyono, Joko
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/43165
• Institution: Institut Teknologi Bandung

Lithium batteries are very popular batteries these past few years. These are transition periods where fossil energy is gradually being replaced by renewable energy and batteries play an important role here as energy storage. One of battery technology that very suitable for storing electrical energy is lithium - ion. Lithium batteries are expensive batteries, so a lot of research aimed to examine the degradation in performance of these lithium-ion batteries. Capacity degradation as a result of the charge effect and its effect on internal resistance are the aim of this study. This research uses a commercial battery and the charging method uses Constant Current – Constant Voltage (CC – CV). This study will also analyze the increase in resistances that occur due to the CC-CV charging method. In the first cycle, the battery's internal resistance is 157,677 mOhm. When it has passed 200 cycles of internal resistance degradates to 162,684 mOhm. This research also predicts battery life using Support Vector Machines (SVM) and Random Forest. By using the Random Forest method generates MSE: 584.925, RMSE: 24.1485, R2: 0.675, Mape: 0.74%. By using the SVM method generates: MSE: 435.381, RMSE: 20.866, R2: 0.758, Mape: 0.65%.