Optimization and characterization of bio-oil produced from cotton shell by flash pyrolysis using artificial neural network

In this work, the effects of pyrolysis parameters such as pyrolysis temperature, feedstock particle size, and sweep gas flow rate on the pyrolysis distributions of cotton shell were tested to identify the optimum bio-oil production conditions. The experiments were conducted in a fluidized-bed reactor with nitrogen atmosphere. The efficient Artificial Neural Network (ANN) was used for modeling and optimizing the process parameters. An ANN model was developed based on the back-propagation learning algorithm to predict the effect of parameters. The experimental data acquired are used for training the network. The results showed that the network yields a maximum bio-oil with minimum coefficient of variance (COV) and root mean square values. The results confirmed that the use of an ANN analysis for pyrolysis distribution is quite suitable. Using the ANN technique, the optimal conditions were found to be at the temperature of 450 degrees C, particle size of 0.8 mm, sweep gas flow rate of 1.75 m(3)/h with 52.2 wt% of bio-oil yield. Furthermore, the liquid product was analyzed for physical, elemental, and chemical composition using Fourier transform infrared (FTIR) spectroscopy and gas chromatography (GC).