Modelling of industrial thermal cracking furnaces via functional-link artificial neural networks

The main thrust of this research is to investigate the feasibility of the use of artificial neural networks for modelling an industrial thermal cracking furnace. The conventional backpropagation network is enhanced by adding a number of functional units to the input layer. This technique considerably extends a network´s capability for representing complex nonlinear relations and makes it possible to predict simultaneously the pyrolysis product distribution and the pyrolysis kinetic severity function (KSF) in an industrial cracking furnace. A very good agreement is obtained between the network model prediction results and actual operational data