Use of neural network for modeling of non-linear process integration technology in chemical engineering

The topology of an industrial petrol refinery complex (PRC) has been formed as a non-linear function by neural network. Starting from the topology of the integration technology’ various choices of production and consumption rates were evaluated both technically and economically on the basis of non-linear material balances. The calculation of the weights of the units of PRC and the determination of the dependence of these units on each other have been performed according to the market conditions. The pertinent equation and the optimal parameters are presented.