Prediction of Effluent Compositions for Fixed-Bed Coal Gasifier Using Artificial Neural Network

The use of an artificial neural network for predicting the effluent compositions in coal gasification was investigated. A mathematical model of coal gasification was developed and simulated to find the effluent compositions for different process conditions. The effluent compositions predicted from the model compare well with our experimental results. The artificial neural network was trained with the simulation data and gave good performance in predicting the effluent compositions. The effect of various parameters like temperature, pressure and steam/coal ratio was studied. Each factor was varied to three levels. At each combination the composition of product gases was predicted by simulation. Then from the above data points, an appropriate Artificial Neural Network (ANN) model were fitted. A three layered feed-forward back-propagation model with 3, 4 and 5 neurons in the layers were fitted. Log-sig function was used in each layer. The model was found to have high R2 (>0.99) value and it had good predicting capability. The results showed that when sufficient amount of training data is available, a neural network model can be developed to predict the effluent compositions of coal gasification with good accuracy and robustness