Thermal Diffusivity of Hydrocarbons and Aromatics and Artificial Neural Network Predicting Model

A feed forward artificial neural network (FFANN) was used to correlate 956 experimentally measured thermal diffusivity data of hydrocarbons and aromatics. In this regards, a networkconsisted of one input, one hidden and one output layers was used to correlate the thermal diffusivity of the hydrocarbon and aromatics. The obtained results revealed that using FFANNmodel contained 19 neurons in the hidden layer and tansig and purelin as transfer functions in hidden and output layers, respectively, leads to overall average absolute relative deviation percent (AARD%) of 0.73 %. Generally, obtained result showed that the proposed ANN model was ableto well correlate the thermal diffusivity of hydrocarbons and aromatics

Keywords:

Thermal diffusivity , Artificial neural network , Hydrocarbons , Aromatics , Prediction