Simulation of the distillation column using artificial neural networks

• Autorzy: Szulc T. , Staszak M.
• Języki publikacji: EN

Abstrakty

EN

Traditional simulation of the dynamics of the distillation process in the rectification column is a time consuming task due to the size of the mathematical models describing the process. The mathematical modeling can be used in the on-line application only in the case when the solution is obtained faster than the real course time of the process. There exist several methods of reducing the computational effort while solving the distillation column model e.g. simplifying the physicochemical or hydrodynamic description, collocation methods or the blackbox models. In this work we propose the simulation of the transient states using the artificial neural network trained using the data from a traditional, mathematical model based on the physicochemical relations. In general, the algorithm of training is based on providing the input values from several preceding times to the input neurons and appropriate adjusting neuron synaptic weights taking into account the difference between the answer obtained at the network output layer and the demanded values. The verification is made using the measurements of the methanol-water mixture distilled in the batch distillation column at the Synteza Sp. z o.o. pharmaceutical factory.

Słowa kluczowe

EN

distillation; simulation

PL

destylacja; kolumna destylacyjna; symulacja; sztuczne sieci neuronowe

• Wydawca: West Pomeranian University of Technology. Publishing House
• Czasopismo: Polish Journal of Chemical Technology
• Rocznik: 2006
• Tom: Vol. 8, nr 4
• Strony: 38--41
• Opis fizyczny: Bibliogr. 4 poz., rys., tab.

Twórcy

autor

Szulc T.

autor

Staszak M.

• Institute of Chemical Technology and Engineering, Poznań University of Technology, pl. M. Skłodowskiej-Curie 2, 60-965 Poznań,

Bibliografia

• (1) Cho Y. S., Joseph B.: Reduced-Order Steady State and Dynamic Models for Separation Processes. American Institute of Chemical Engineers Journal, 1983a, 29, 261.
• (2) Cho Y. S., Joseph B.: Reduced-Order Models for Separation Columns - Application to columns with multiple feeds and sidestreams. Computers and Chemical Engineering, 1983b, 8, 81.
• (3) Alejski K., Duprat F.: Dynamic simulation of the multicomponent reactive distillation, Chemical Engineering Science, 1996, 51, 4237.
• (4) Alejski K.: Computation of the reacting distillation column using a liquid mixing model on plates. Computers and Chemical Engineering, 1991, 15, 313.