Modelling of biodiesel production from transesterification process of sandbox (Hura crepitans L.) seed oil: performance comparison of artificial neural networks (ANN) and adaptive neuro-fuzzy inference system (ANFIS) (original) (raw)
Related papers
American Journal of Applied Chemistry
Biodiesel has been referred to as a perfect substitute for diesel fuel due to its numerous promising properties. They are renewable, clean, increases energy security, improves the environment and air quality and also provides some good safety benefits. This study is focused on the investigation of the use of natural heterogeneous catalysts for production of biodiesel from jansa seed oil, as well as the implementation of artificial neural network (ANN) for the prediction of biofuel yield and process parameters. The biodiesel was produced through transesterification reaction by reacting jansa seed oil (FFA) with methanol (alcohol) to yield methyl ester. Waste periwinkle shell was prepared in 3 different forms; raw, calcined and acidified. The percentage yield of the methyl ester obtained were calculated and tabulated. The process parameters considered were methanol-oil mole ratio, catalyst concentration, agitation speed, reaction temperature and reaction time. The results of this research work revealed that the calcined periwinkle shell catalyst produced higher yield of biodiesel, compared to the yield obtained from the raw and acidified catalyzed process. The properties of the fatty acid methyl esters were within the standard range. The experimental and predicted yield were marginally the same. Hence, the model accurately predicted the yield with acceptable coefficient of determination and low mean squared error (MSE). The results demonstrate the flexibility of ANN model and the improvement of the model in terms of performance prediction when solving problems with stochastic dataset, especially the transesterification of biodiesel.
This work focused on the application of adaptive neuro-fuzzy inference system (ANFIS) and response surface methodology (RSM) as predictive tools for production of fatty acid methyl esters (FAME) from yellow oleander (Thevetia peruviana) seed oil. Two-step transesterification method was adopted, in the first step, the high free fatty acid (FFA) content of the oil was reduced to <1% by treating it with ferric sulfate in the presence of methanol. While in the second step, the pretreated oil was converted to FAME by reacting it with methanol using sodium methoxide as catalyst. To model the second step, central composite design was employed to study the effect of catalyst loading (1-2 wt.%), methanol/oil molar ratio (6:1-12:1) and time (20-60 min) on the T. peruviana methyl esters (TPME) yield. The reduction of FFA of the oil to 0.65 ± 0.05 wt.% was realized using ferric sulfate of 3 wt.%, methanol/FFA molar ratio of 9:1 and reaction time of 40 min. The model developed for the transesterification process by ANFIS (coefficient of determination, R 2 = 0.9999, standard error of prediction, SEP = 0.07 and mean absolute percentage deviation, MAPD = 0.05%) was significantly better than that of RSM (R 2 = 0.9670, SEP = 1.55 and MAPD = 0.84%) in terms of accuracy of the predicted TPME yield. For maximum TPME yield, the transesterification process input variables were optimized using genetic algorithm (GA) coupled with the ANFIS model and RSM optimization tool. TPME yield of 99.8 wt.% could be obtained with the combination of 0.79 w/v catalyst loading, 12.5:1 methanol/oil molar ratio and time of 58.2 min using ANFIS-GA in comparison to TPME yield of 98.8 wt.% using RSM. The TPME structure was characterized using Fourier transform infra-red (FT-IR) spectroscopy. The results of this work established the superiority of predictive capability of ANFIS over RSM.
Renewable Energy, 2013
In this study, transesterification of soybean oil to biodiesel using KOH in different process conditions were studied. The investigated conditions were the molar ratio of methanol/oil, catalyst amount and reaction temperature. Optimal conditions were found to be methanol/oil molar ratio, 9:1; catalyst amount, 1 wt%; reaction temperature, 60 C. Biodiesel yield for these conditions was obtained 93.2% in 1 h. In addition, the artificial neural network has been applied to estimate the biodiesel yield. The multilayer feed forward neural network with three inputs and one output has been trained with different algorithms and different numbers of neurons in the hidden layer. The accuracy of the proposed model was found to agree nearly with the experimental results over a wide range of experimental conditions. The results clearly depict that the neural network is a powerful tool to estimate the reaction rate and the designed neural network can be used instead of approximate and complex analytical equations.
Modelling of Nicotiana Tabacum L. Oil Biodiesel Production: Comparison of ANN and ANFIS
Frontiers in Energy Research, 2021
Among the modern computational techniques, Artificial Neural Network (ANN) and Adaptive Neuro-Fuzzy Inference System (ANFIS) are preferred because of their ability to deal with non-linear modelling and complex stochastic dataset. Nondeterministic models involve some computational complexities while solving real-life problems but would always produce better outcomes. For the first time, this study utilized the ANN and ANFIS models for modelling tobacco seed oil methyl ester (TSOME) production from underutilized tobacco seeds in the tropics. The dataset for the models was obtained from an earlier study which focused on design of the experiment on TSOME production. This study is an an exposition of the influence of transesterification parameters such as reaction duration (T), methanol/oil molar ratio (M:O), and catalyst dosage on the TSOME/biodiesel yield. A multi-layer ANN model with ten hidden layers was trained to simulate the methanolysis process. The ANFIS approach was further imp...
Environmental Progress & Sustainable Energy, 2020
The conversion of sorrel oil to biodiesel through transesterification was conducted in the presence of calcined kola nut husk pod ash as a base heterogeneous catalyst. Thus, to predict the biodiesel production yield, two models based on neural intelligence and neuro-fuzzy techniques were established. The predictive capability and accuracy of the models were compared using various statistics. The neuro-fuzzy-based model (adaptive neuro-fuzzy inference system-ANFIS) obtained for the transesterification process had a lower (0.3247) mean relative percent deviation (MRPD) and a higher coefficient of determination-R 2 (0.9991) compared to the neural intelligence-based model (artificial neural network-ANN) with MRPD of 0.42% and R 2 of 0.9971. Also, the models developed were coupled with genetic This article is protected by copyright. All rights reserved. algorithm (GA) in order to maximize the sorrel oil biodiesel (SOB) yield at optimum values of the process input parameters. SOB yield of > 99.0 wt.% was obtained when both developed models were subjected to optimization. The results of the process modeling confirm that neurofuzzy model performed slightly better than neural intelligence model. The sensitivity analysis performed on both models shows that reaction time was the most important input variable while other input variables could not be neglected. The characteristics of the synthesized SOB demonstrate that it satisfied the biodiesel standard limits.
ELSEVIER, 2020
The biodiesel production from waste soybean oil (using NaOH and KOH catalysts independently) was investigated in this study. The use of optimization tools (artificial neural network, ANN, and response surface methodology, RSM) for the modelling of the relationship between biodiesel yield and process parameters was carried out. The variables employed in the experimental design of biodiesel yields were methanol-oil mole ratio (6 –12), catalyst concentration (0.7 –1.7 wt/wt%), reaction temperature (48 –62 °C) and reaction time (50 –90 min). Also, the usefulness of both the RSM and ANN tools in the accurate prediction of the regression models were revealed, with values of R-sq being 0.93 and 0.98 for RSM and ANN respectively.
This study investigates the applicability of adaptive neuro-fuzzy inference system (ANFIS) and artificial neural network (ANN) approaches for modeling the fatty acid methyl esters (FAMEs) property including kinematic viscosity at various temperatures and the volume fractions of biodiesel. An experimental database of kinematic viscosity of pure biodiesel was used for developing of models, where the input variables in the network were the temperature, the number of carbon atoms (NC) and the number of hydrogen atoms (NH) of the composition of methyl esters (C8:0, C10:0, C12:0, C14:0, C16:0, C16:1, C18:0, C18:1, C18:2, C18:3, C20:0, C20:1, C22:0, C22:1, C24:0 were considered as input variables on the ANFIS and ANN. Moreover, the models are divided into saturated species from C8:0 to C24:0 and unsaturated species, from C16:1 to C22:1. The model results were compared with experimental ones for determining the accuracy of the ANFIS and ANN predictions. The developed model produced idealized results and was found to be useful for predicting the kinematic viscosity of biodiesel blends with a limited number of available data. Moreover, the results suggest that the ANFIS approach can be used successfully for predicting the kinematic viscosity of biodiesel blends at various volume fractions and temperature compared to ANN approach.
Prediction of biodiesel production from palm fatty acid distillate using artificial neural network
2018
Sustainable and low cost feedstock, i.e. palm fatty acid distillate (PFAD) is one of potential source to be utilized in biodiesel industry. Non-catalytic biodiesel production using supercritical technology is considered as a promising method to treat feedstock containing high free fatty acid (FFA) such as PFAD owing to no catalyst needed and no sensitive to the presence of FFA. However, there has been no previous study to predict biodiesel production from PFAD. In this study, artificial neural network was used to predict biodiesel yield from PFAD for the first time. The experimental data of biodiesel yield conducted by varying 3 input factors, i.e. temperature, oil-to-methanol molar ratio, and residence time were used to elucidate artificial neural network model in order to predict biodiesel yield. The objective this study is to assess how accurately this artificial neural network model to predict biodiesel yield from PFAD conducted under supercritical methanol condition. The result...
Yield Optimization Using Artificial Neural Networks in Biodiesel Production from Soybean Oil
Revista de Chimie
Biodiesel plays an important role in reducing the dependency of petroleum fuels and reduce environmental pollution. Biodiesel has attracted attention as a renewable, non-toxic, and biodegradable fuel. In the past years, researchers have expanded their work to new methods of obtaining biodiesel. In this work, biodiesel production using soybean oil from the industry is described. Biodiesel was further analyzed and compared with the EN biodiesel specifications. The characterization of biodiesel was performed in order to obtain density, viscosity and flash point. Moreover, the study was focused in optimized the biodiesel yield, obtain from soybean oil using Artificial Neural Networks (ANN). The variable parameters were molar ratio between methanol and oil, reaction temperature and catalyst quantity. The paper concludes that the ANN can be successfully used to optimize the biodiesel yield.
Application of Neural Network for Estimating Properties of Diesel–Biodieselblends
International Journal of Computers and Applications, 2010
Soybean oil was transesterified with methanol in the presence of alkaline catalyst to produce methyl esters commonly known as biodiesel. Biodiesel and diesel blends were prepared and tested in laboratory for flash point, fire point, viscosity and density. Seven neural network architectures, three training algorithms along with ten different sets of weight and biases were examined to predict the above-mentioned properties of diesel and biodiesel blends. The best suited neural network and training algorithm were selected and further generalized to improve its performance by using early stopping technique. The results showed that the neural network having an architecture 2-7-4 with Levernberg-Marquardt algorithm gave the best estimate for the properties of diesel-biodiesel blends.