BATTERY STATE OF CHARGE ESTIMATION MODEL USING VOLTAGE DROP

• Main Author: NURUL HIKMAH (nim : 13313024) - MUHAMMAD AKMALUL IHSAN (nim : 13313052), NAHDIA
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/29512
• Institution: Institut Teknologi Bandung

<br /> <br /> For the past two decades, global climate change have caused negative impacts on the environment. This fact has pushed the advancement of green technology, which utilizes renewable resources as an energy source. One of the most crucial part in this research is the energy storage. Batteries are the most conventional energy storage systems, used in various applications from hand phones to electrical vehicles. <br /> <br /> The amount of charge in a battery determines the life span of the device that uses said battery. An indicator of the amount of charge left in a battery is the SOC (State of Charge). The objective of this final project is to create a model that can estimate SOC values accurately and in real-time, and also define the suitable range to use said model. Model construction starts by introducing a new variable called unit time voltage change (V’). This research is then expanded to see whether an SOC model of a battery can be used to estimate the SOC of other batteries that are produced in the same batch. <br /> <br /> From 3 life cycle data an SOC model is derived. Results from this SOC model is compared to the values from SOC estimation using Coulomb counting method (experimental). This research is then developed to see if the SOC model derived from battery 1 can be used to estimate the SOC from other batteries that are produced in the same batch. Model accuracy evaluation is quantified by RMSE (Root Mean Square Error) and coefficient of determination (R2) values. <br /> <br /> The SOC model constructed from battery 1 can estimate the real value of SOC accurately, shown by the R2 score 0.953 and the RMSE value 0.043. Battery 1 SOC model is then used to estimate the SOC of other batteries. The robustness of the model is proved with the R2 score stayed above 0.937 and RMSE value stayed below 0.052.Model constructed from battery 1 can estimate the SOC of all batteries with a maximum error 3.74%. Small error values in estimating SOC values occur when the voltage is in 3.55 V – 3.95 V, where those voltage values responds to SOC values of 40% - 90%. <br />