Advanced predictive modeling of biodiesel yield using artificial neural networks for Karanja oil transesterification

Biodiesel has become more popular as a diesel substitute in recent years due to environmental concerns and the depletion of vital resources like coal and petroleum. Because of the finite supply of petroleum, concerns about global warming, and annual rises in petroleum prices, researchers are currently searching for new, affordable energy sources. The best substitute for diesel fuel is biodiesel, which is becoming more and more well-liked because of its benefits and capacity to protect the environment. The research explored the prediction of biodiesel synthesis and included transesterification experiments involving Karanja oil. In order to forecast the yield of biodiesel, an artificial neural network model was designed. Using an artificial neural network, four process parameters were evaluated during the procedure: the Oil-to-methanol molar ratio (40–50%), proportion of the catalyst (1–2 wt.%), power (300–400 watts) and duration (4-6 minutes). The experiment produced biodiesel with a 99.98% yield using a catalyst (1.5 weight percent), duration (5 minutes), methanol (50%), and power (300 watts). An ANN was trained using the Levenberg-Marquardt algorithm. The estimations and the outcomes of the experiment were contrasted. The ANN model's R-squared value was 0.9487.