Modelling of a complex, heterogeneous process of sodium nitrate (III) inversion with the nitric acid (V) was carried out. The Artificial Neural Network was used for the process kinetics' numerical evaluation. The own set of experimental data was the basis for the Neural Network's training and testing. The optimal Neural Network's structure corresponds to: 4 inputs, 2 hidden layer (with adequately: 10 and 6 neurones) and 2 output neurones. The Neural Network's model correctly describes the inversion process kinetics in the following parameters' ranges: c(0NaNO2) = 0.25-0.5 mol/dm3, c(0NaNO3) = 1.5 - 3.0 mol/dm3, T = 303 - 343 K, Re(m) = 8500.
2
Dostęp do pełnego tekstu na zewnętrznej witrynie WWW