Fuzzy logic model for degumming and bleaching troubleshooting in palm oil refining

Abstract

Failures at degumming and bleaching process in palm oil refining affect plant performance, production delay, and loss to the company. In current practice, troubleshooting in these process relies on human knowledge and trial and error method which can caused other failures and time consuming. In this study, fuzzy logic model was developed for troubleshooting degumming and bleaching process using Mamdani approached. Operating conditions at bleacher vessel, ejector, condenser, hotwell and cooling tower system were selected as input and output variables in the fuzzy logic model development. Qualitative and numerical data were collected in this study by interviewing plant workers, observing distributed control system (DCS), reviewing standard operating procedure (SOP), operation checklist and plant manuals. The relationship between input and output variables were described by fuzzy membership function and fuzzy rules. Centre of gravity method (COG) method was used for defuzzification. The proposed model was tested with real plant data and the model was shown successfully perform troubleshooting task and suggest necessary action. Therefore, this model has the potential to employed by operator and inexperience workers in palm oil industry.