Quality prediction modeling of palm oil refining plant in Malaysia using artificial neural network models

Abstract

Malaysia is currently one of the biggest producers and exporters of palm oil and palm oil products. In the growth of palm oil industry in Malaysia, quality of the refined oil is a major concern where off-specification products will be rejected thus causing a great loss in profit. In this paper, predictive modeling of refined palm oil quality in one palm oil refining plant in Malaysia is proposed for online quality monitoring purposes. The color of the crude oil, Free Fatty acid (FFA) content, bleaching earth dosage, citric acid dosage, activated carbon dosage, deodorizer pressure and deodorizer temperature were studied in this paper. The industrial palm oil refinery data were used as input and output to the Artificial Neural Network (ANN) model. Various trials were examined for training all three ANN models using number of nodes in the hidden layer varying from 10 to 25. All three models were trained and tested reasonably well to predict FFA content, red and yellow color quality of the refined palm oil efficiently with small error. Therefore, the models can be further implemented in palm oil refinery plant as online prediction system.