Estimation Quality Monitoring Glycerol Esterification Process with IR Sensors Using K Nearest Neighbours Classification

Abstract

The commercial synthesis of fatty acid esters of glycerol is important because it can be used for other derivative production varieties. This research aims to construct the quality monitoring system for esterification condition faster and more efficient for the production of esterification glycerol. The monitoring systems were based on the measurement parameters from two inputs LED mid IR 3,4 and 5,5 μm sensors that using the data acquisition with computer database via USB 2.0 using Arduino Leonardo microcontroller and classifying the esterification quality condition using the classification method K-Nearest Neighborhood (KNN) The purpose of KNN method is to classify the variations of parameter inputs from the LED mid IR sensors in quality monitoring. In this research the condition of esterification was divided into three conditions: not good, fair,good., these classification was trained and tested in Orange Software for data mining using receiver operating characteristic (ROC) curve that is a graphical plot that illustrates the excellent performance of a classifier system for esterification condition with AUC . In application for quality monitoring, the influence of various parameters such as temperature set in the reactor has relation to the quality of product. By using this system, we obtained the optimum process conditions is 200oC and time needed for the process was 200 minutes.