Title :
Application of near infrared spectroscopy for assessment of free fatty acid in palm oil
Author :
Bt Abdull Rani, Siti Nurhidayah Naqiah ; Binti Abdul Rahim, Herlina
Author_Institution :
Control & Instrum. Eng. Dept., UTM, Skudai, Malaysia
Abstract :
This paper propose an experiment to determine compatibility of near infrared (NIR) application in crude palm oil (CPO) quality measurement, considering only the Free Fatty Acid (FFA) that can be made as open source software. 100 samples were taken from Felda Johor Bulkers, Terminal 1 in 7 consecutive days of experiment. FFA reading from conventional wet chemical analysis is used as the reference and validation for NIR method. Partial Least Square Regression (PLSR) method is used for the linear analysis, generated from MATLAB software. However, the finding from this experiment not really support the application at which the accuracy percentage acquire is only 48.12%. Further investigation will be used Neural Network for give a better accuracy compared with PLSR.
Keywords :
crude oil; essential oils; food processing industry; infrared spectroscopy; least squares approximations; mathematics computing; neural nets; product quality; production engineering computing; regression analysis; CPO quality measurement; FFA; Felda Johor Bulkers; MATLAB software; NIR application; PLSR method; conventional wet chemical analysis; crude palm oil quality measurement; free fatty acid assessment; linear analysis; near infrared spectroscopy; neural network; open source software; partial least square regression method; Accuracy; Length measurement; Software; FFA; NIRS; PLSR; palm oil; sgolay;
Conference_Titel :
Open Systems (ICOS), 2013 IEEE Conference on
Conference_Location :
Kuching
Print_ISBN :
978-1-4799-3152-1
DOI :
10.1109/ICOS.2013.6735044
