Composition Prediction of Debutanizer Column using Neural Network

In oil refining industries, debutanizer column is one of the important unit
operations. Debutanizer column is the main column used to produce the main
product in oil refinery process. The online composition prediction of top and bottom
product of debutanizer column using neural network will be an aid to increase
product quality monitoring in oil refining industry. In this work, a single dynamic
neural network model is used in order to achieve the objective which is to generate
composition prediction online of the top and bottom product of debutanizer column.
Neural network is a computing system with several of simple and highly
interconnected processing elements that will process information using their dynamic
state response to external inputs. It is a software based sensor method or known as
“soft sensor” which is a helpful technology that utilizes software techniques to infer
the value of important but difficult-to-measure process variables from available
process variables which are requisite from physical sensor observation or lab
measurements. The neural network development and equation based model for ibutane,
i-pentane, n-butane, n-pentane and propane has been obtained. Then, these
results will be compared with proportional integral derivatives (PID) controller
design to show its supremacy over this method.