loading page

Using major components of biodiesel: Evaluating the applicability of different machine learning for predicting the CFPP
  • +1
  • wenchao Wang,
  • Huini Qi,
  • Fashe Li,
  • Hua Wang
wenchao Wang
Author Profile
Fashe Li

Corresponding Author:[email protected]

Author Profile

Abstract

Prediction of CFPP for biodiesel based on fatty acid methyl esters composition is the focus of the presented research. A total of 90 fuel samples were derived from the literature and an additional 10 fuel samples were obtained through experimentation in order to provide an initial data set for predictive model development and training. The adaptability of RBFNN in predicting the CFPP of biodiesel is discussed, and the prediction performance compared with MLR and BPNN. When using MLR, BPNN and RBFNN for predicting CFPP of biodiesel, results demonstrated a R2 of 0.9273, 0.9990 and 0.9999, respectively. By subsequently calculating the performance evaluation index of the model, the RBFNN model's RMSE, NRMSE and SSE are the closest to 0 compared to the other two techniques. Based on the outcome of this study, these techniques can be effective tools for predicting CFPP of biodiesels, with RBFNN showing the most promise.