Comparing refined palm oil product quality predictor in palm oil refineries using partial correlation analysis and partial least square.

Abstract

As a part of continuous improvement in line with Industry 4.0 targets, smart quality prediction tools have been developed in order to forecast the quality of the refined, bleached and deodorized palm oil (RBDPO). RBDPO quality forecasting model development began with data collection, followed by a pre-processing stage to acquire the optimum sampling time and the processing time of the refining process. Using the pre-processed data, the predictor coefficients are then developed using Partial Correlation Analysis (PCorrA) and Partial Least Square (PLS) algorithms with the help of MATLAB programming software, and the forecasted data plotted together with the actual real time data in control charts to assess the refining process performance of Lahad Datu Edible Oils Sdn. Bhd. (LDEO). The sampling frequency is reduced by 75 % as product sampling time is set at every four hours. The residence time selected at eight hours. Through mean squared error (MSE) computations, PCorrA shows consistently low MSE readings of 0, 0, 0.0036 and 0.0450 for free fatty acid (FFA), moisture (MOIST), iodine value (IV) and COLOR. The RBDPO quality results from 100 crude palm oil (CPO) tankers show PCorrA able to predict RBDPO quality. PCorrA is selected as the better forecasting algorithm against PLS.