Improved lithium-ion battery model for photovoltaic applications based on comparative analysis and experimental tests

Lithium-ion (Li-ion) batteries have improved crucially and are widely regarded as a vital component in the expansion of renewable energy sources. Actually, they are still undergoing significant development, owing to their progressive spread from wearable electronics to more advanced fields such as electric vehicles and smart grids. Therefore, battery modeling facilitates forecasting, as well as efficient operation and precise energy predictions. Furthermore, a vast number of models for a wide range of applications have been manufactured globally over time. In light of this, 10 battery models, covering electrochemistry, mathematical, and electric models, are examined in depth, taking the parameter optimization into consideration with the Rosenbrock pattern research technique. The suggested models will be compared using quantitative and qualitative analysis as well as other statistical error metrics such as the mean bias error and the root mean square error. In all the models discussed, the overall impact of voltage prediction is rigorously examined and compared with charging and discharging modes during a real photovoltaic system. The results reveal that the battery voltage predicted by the Nernst model was highly accurate; in charge and discharge modes, the bulk of the predictions had near-zero relative error and a maximum error value of approximately 6 mV. Current research offers insights into concepts that may be used to motivate solar system researchers as well as advise the optimal battery model choices for future photovoltaic storage advancements. Highlights Experimental study of off-grid PV system with lithium-ion battery storage is done. Modeling and simulation of ten battery models have been achieved. Correlation using the Rosenrock pattern research algorithm is done. Comparative study and statistical analysis are discussed.