Abstract:

This paper presents the research and application of backpropagation algorithm of artificial neural network to create models that predict flooding of distillation column in the petroleum industry. The Matlab Simulink software has been used for modelling on the basis of the equations describing the process kinetics of the column, the survey of mathematical equations and evaluate the possibility of flooding. The data received from the modeling forms the basis for building neural network models. Results achieved in the modeling can be used for forecasting floods, as well as online measurement of flooding in distillation columns, which can assist the warning and prevent the incident occurred due to flooding of the column. As a result, the product quality, operation and safety of the distillation column in petroleum processing industry can be improved