Using Artificial Bee Colony and Dragonfly Algorithms to Improve the Accuracy of Parameter Estimation of Li- FePO4 Battery Cell

Studies on li-ion batteries, which are increasingly used in various technological fields, are spreading to different areas. One of these study subjects is to determine the battery parameters with high accuracy and transfer them to the test and simulation system. In a study conducted for a similar purpose, the A123 ANR26650 Li-FePO4 battery cell has been discharged at room conditions with 2000 mA constant current. The real terminal voltage has been obtained by analyzing the obtained data. Mathematical expressions of internal resistance and terminal voltage are defined. The terminal voltage is characterized as a third-order function dependent on the state of charge. Simulink/Matlab, curve fitting, artificial bee colony algorithm, and dragonfly algorithm are used to forecast terminal voltage. While comparing the methods, the Simulink model has not been included in the comparison because it has been the ideal model. As a result, the most consistent method has been determined by using different statistical methods such as the actual error values, the sum of the squares of the errors, and the mean value of the sum of the squares of the errors. Artificial bee colony algorithm has performed best with an mean value of the sum of the squares of the errors of 1.1822e-05. It has emerged that dragonfly algorithm needs to be developed and a new field of study has been opened.