Machine Learning-Assisted Electrode Material Fabrication and Electrochemical Efficiency Prediction and Validation of PANI-Ni/Co Hydroxide Nanocomposites

Abstract

Electrode materials play a critical role in the charge storage mechanism in supercapacitor devices. Hence, predicting the performance, namely, the specific capacitance ( C sp ), rate capability, and cyclic stability of these electrodes, is crucial for supercapacitor applications. In this report, a data-centric machine-learning (ML) approach was implemented to study the effect of polyaniline (PANI) on the electrochemical behavior of NiCo(OH) 2 . ML models were used to predict the combinational C sp -grade and C sp -value of composite electrode materials. The lowest relative percentage error of 1.04% for 3 PNC is observed between predicted and experimental C sp -values. The ML-predicted rate capability of 3 PNC strongly aligned with the experimental data when the current density is increased 4-fold and revealed better rate capability. The experimental C sp retention of 82.4% after 10000 cycles closely matched the predicted retention of 88.56%, with a percentage error of 6.92%. Hence, the ML may be implemented to predict the performance of other electrode materials for energy storage applications.