Abstract (EN):
A discussion covers how feedforward artificial neural networks (FANN) are utilized to model rotating disc contactors for deasphalting and for furfural extraction, a valuable approach in solving complex problems in the refining business; situations where the phenomenological models are practically unknown due to the process complexity and the difficulty in developing satisfactory thermodynamic models for the complex mixtures of pseudocomponents involved, i.e., propane deasphalting process and the furfural refining process; features of a FANN, e.g., interconnected non-linear processing elements called nodes; FANN created based on data corresponding to the operational conditions for three different lubricating oil cuts; important variables in the propane deasphalting process, i.e., feedstock quality, solvent, and deasphalting temperatures; and important variables for the solvent refining process, i.e., crude source, distillation cut, solvent temperature, solvent dosage, and extract recycle.
Language:
English
Type (Professor's evaluation):
Scientific
No. of pages:
5