Optimizing Biodiesel Yield and Fuel Properties from Waste Avocado Oil: A Comparative Study of RSM and ANFIS Learning Models
نویسندگان
1 Department of Electrical Electronics Engineering, Kent State University, USA
2 Department of Chemical Engineering, Federal University of Technology, Owerri, Nigeria
3 Department of Chemical Engineering, Federal university of Technology, Minna, Niger state, Nigeria
4 Department of Chemical Engineering, Ladoke Akintola University of Technology, Nigeria
5 Department of Environmental Design, University of Lagos, Nigeria
6 Department of Mechatronics Engineering and Robotics, MIREA Russian Technological University, Moscow, Russia
7 Department of Mechanical Engineering, Federal University of Petroleum Resources, Effurun, Nigeria
8 Federal University of Technology Owerri, Nigeria
9 Department of Oil and Gas Transport and Refinery Operation Engineering, Kazan National Research Technological University, Kazan, Russian
10 Department of Petroleum and Gas Engineering, University of Lagos, Nigeria
11 Department of Industrial Environmental Engineering, Universiti Teknologi PETRONAS, Malaysia
12 Department of Chemical Engineering, Federal University of Technology, Minna, Nigeria
13 Department of Industrial Environmental Engineering, Universiti Teknologi PETRONAS, Malaysia
14 Department of Chemical Engineering, Lagos state University, Nigeria
15 Department of Information Technology, Federal university of Technology, Owerri, Nigeria
16 Department of Chemical Engineering, Federal University of Technology Minna, Nigeria
doi
10.48309/pcbr.2025.495665.1389چکیده
The quest for alternative energy sources, driven by the challenges of fossil fuels, has led to the development of biofuels. This study focuses on producing biodiesel from waste avocado oil through transesterification. Initially, oil was extracted from the peels and seed of avocado pear using an extraction technique. The extracted oil was then pre-treated with methanol and sulfuric acid (H₂SO₄) to reduce its free fatty acid content to less than 1.0 wt%. This study compares two expert systems, Adaptive Neuro-Fuzzy Inference System (ANFIS), and Response Surface Methodology (RSM), for modeling and optimizing biodiesel production from avocado oil. The performance of these optimization tools was evaluated using statistical indices. The results showed that ANFIS outperformed RSM with a low error value, the Standard Error of Prediction (SEP)=0.7653, the Mean Absolute Error (MAE)=0.1413, the Root Mean Squared Error (RMSE)=0.4103, the Average Absolute Deviation (AAD)=0.2955%, the Mean Squared Error (MSE)=0.1683, and a high coefficient of determination (R² = 0.9976). Both models predicted high biodiesel yields (>85%), with ANFIS achieving a slightly higher yield (88.21%) compared to RSM (86.20%). The properties of the biodiesel produced under optimized conditions were compared with American Society for Testing and Materials (ASTM) D6751 and European Norm (EN) 14214 standards and were found to be within acceptable limits, indicating the fuel’s suitability.
